indusagi-coding-agent 0.1.37 → 0.1.39

package/docs/PI_MONO_MIT_REMOVAL_GUIDE.md

@@ -0,0 +1,2132 @@
+# PI Mono MIT Removal Guide for `indusagi-screth`
+
+
+
+> **Important:** This is a technical provenance-cleanup guide, **not legal advice**. If you want to make a strong public claim like _"`indusagi-screth` contains no PI Mono derived code"_, you should still do a final legal/provenance review.
+
+
+
+## Goal
+
+
+
+If the goal is to make `indusagi-screth` **PI Mono MIT-free**, the safe standard is:ß
+
+
+
+1. No copied PI Mono source files remain.
+
+2. No adapted/derived PI Mono files remain unless you are willing to keep MIT attribution.
+
+3. Third-party vendor assets are either:
+
+   - replaced by direct upstream package installation/build steps, or
+
+   - kept with their own proper attribution, independent of PI Mono.
+
+4. Documentation, package metadata, and marketing text do not imply PI Mono-derived implementation if that is no longer true.
+
+
+
+## What the current comparison shows
+
+
+
+Compared repos:
+
+- `@pi-mono/packages/coding-agent/`
+
+- `@indusagi-screth/`
+
+
+
+Current overlap snapshot:
+
+- common relative paths: `121`
+
+- exact identical files: `10`
+
+- same path but different content: `111`
+
+
+
+That means `indusagi-screth` is **not** currently in a position where it is safe to say _"PI Mono copyright does not apply"_. There is still enough exact and likely-derived overlap that a cleanup pass is required.
+
+
+
+---
+
+
+
+# Priority plan
+
+
+
+## P0 — Rules before touching files
+
+
+
+Before rewriting anything, follow these rules:
+
+
+
+- Do **not** use PI Mono files as copy-paste sources during the rewrite.
+
+- Rewrite from behavior/specification, not from line-by-line adaptation.
+
+- Keep a provenance note for each rewritten file:
+
+  - `rewritten independently`
+
+  - `replaced with upstream vendor asset`
+
+  - `kept and MIT attribution retained`
+
+- If a file is still PI Mono-derived after cleanup, keep MIT attribution for that file/subtree.
+
+
+
+---
+
+
+
+## P1 — Must rewrite first: exact identical files
+
+
+
+These are the highest-risk files because they currently match PI Mono exactly.
+
+
+
+### P1-A. Exact code files to rewrite from scratch
+
+- `src/core/compaction/index.ts` — similarity: 5.7%, difference: 94.3%, status: DIFFERENT
+
+- `src/modes/index.ts` — similarity: 12.8%, difference: 87.2%, status: DIFFERENT
+
+- `src/modes/interactive/components/diff.ts` — similarity: 10.1%, difference: 89.9%, status: DIFFERENT
+
+- `src/utils/changelog.ts` — similarity: 25.7%, difference: 74.3%, status: DIFFERENT
+
+- `src/utils/mime.ts` — similarity: 16.9%, difference: 83.1%, status: DIFFERENT
+
+- `src/utils/photon.ts` — similarity: 17.4%, difference: 82.6%, status: DIFFERENT
+
+- `src/utils/sleep.ts` — similarity: 18.5%, difference: 81.5%, status: DIFFERENT
+
+
+
+### P1-B. Exact vendor / asset files to replace properly
+
+These should not stay as copied repo assets if the goal is "PI Mono MIT-free provenance".
+
+
+
+- `src/core/export-html/template.html` — similarity: 43.1%, difference: 56.9%, status: DIFFERENT
+
+- `src/core/export-html/vendor/highlight.min.js` — similarity: 68.6%, difference: 31.4%, status: DIFFERENT
+
+- `src/core/export-html/vendor/marked.min.js` — similarity: 7.5%, difference: 92.5%, status: DIFFERENT
+
+
+
+### P1 action
+
+For each file above, do one of:
+
+- full independent rewrite, or
+
+- replace from original upstream package/source with proper direct attribution, not via PI Mono copy.
+
+### P1 status update
+
+P1 has now been executed for these files.
+The original exact-match state is gone for this batch, and the percentages below reflect the **post-rewrite / post-replacement** state.
+
+### P1 review-first plan — minimal change, maximum effect
+
+For this P1 set, the best strategy is **small rewrites with structural impact**.
+
+That means:
+- keep runtime behavior compatible where possible
+- prefer rewriting the smallest exact-match files first
+- change helper structure, export shape, naming model, and control flow
+- avoid fake rewrites like only changing comments, spacing, quote style, or variable names
+
+> **Important:** only changing names, comments, formatting, or coding style is **not enough** for a strong provenance-cleanup story. The code should be recognizably reimplemented, not cosmetically edited.
+
+### P1 recommended execution order
+
+1. `src/core/compaction/index.ts` — similarity: 5.7%, difference: 94.3%, status: DIFFERENT
+2. `src/modes/index.ts` — similarity: 12.8%, difference: 87.2%, status: DIFFERENT
+3. `src/utils/sleep.ts` — similarity: 18.5%, difference: 81.5%, status: DIFFERENT
+4. `src/utils/mime.ts` — similarity: 16.9%, difference: 83.1%, status: DIFFERENT
+5. `src/utils/changelog.ts` — similarity: 25.7%, difference: 74.3%, status: DIFFERENT
+6. `src/utils/photon.ts` — similarity: 17.4%, difference: 82.6%, status: DIFFERENT
+7. `src/modes/interactive/components/diff.ts` — similarity: 10.1%, difference: 89.9%, status: DIFFERENT
+8. `src/core/export-html/template.html` — similarity: 43.1%, difference: 56.9%, status: DIFFERENT
+9. `src/core/export-html/vendor/highlight.min.js` — similarity: 68.6%, difference: 31.4%, status: DIFFERENT
+10. `src/core/export-html/vendor/marked.min.js` — similarity: 7.5%, difference: 92.5%, status: DIFFERENT
+
+### P1 per-file plan
+
+#### `src/core/compaction/index.ts` — similarity: 5.7%, difference: 94.3%, status: DIFFERENT
+
+**Why first:** tiny barrel file, low risk, fast win.
+
+**Minimal but meaningful rewrite:**
+- replace wildcard re-exports with explicit named exports
+- import from child modules first, then re-export
+- optionally add a small grouped export object like `compactionApi`
+
+**Avoid:**
+- changing only the header comment
+- reordering the same `export *` lines
+
+#### `src/modes/index.ts` — similarity: 12.8%, difference: 87.2%, status: DIFFERENT
+
+**Why first:** also tiny and easy to rewrite cleanly.
+
+**Minimal but meaningful rewrite:**
+- stop using the same export arrangement
+- import values/types explicitly and export them in a new grouped shape
+- optionally add a small local `modes` registry object
+
+**Avoid:**
+- same barrel pattern with only reordered lines
+
+#### `src/utils/sleep.ts` — similarity: 18.5%, difference: 81.5%, status: DIFFERENT
+
+**Minimal but meaningful rewrite:**
+- split abort handling into helpers
+- add a local `createAbortError()` helper
+- ensure listener cleanup with `{ once: true }` or explicit removal
+- keep public API compatible
+
+**Suggested rewrite shape:**
+- `createAbortError()`
+- `bindAbortHandler()`
+- `sleep()` orchestrates the promise
+
+#### `src/utils/mime.ts` — similarity: 16.9%, difference: 83.1%, status: DIFFERENT
+
+**Minimal but meaningful rewrite:**
+- separate file reading from MIME validation
+- add a helper like `isSupportedImageMimeType()`
+- optionally add extension fallback logic
+- keep the top-level exported API stable if needed
+
+**Suggested rewrite shape:**
+- `readHeaderBytes()`
+- `detectMimeFromBuffer()`
+- `isSupportedImageMimeType()`
+- `detectSupportedImageMimeTypeFromFile()`
+
+#### `src/utils/changelog.ts` — similarity: 25.7%, difference: 74.3%, status: DIFFERENT
+
+**Minimal but meaningful rewrite:**
+- switch from line-by-line state tracking to section splitting
+- isolate version parsing and comparison logic
+- keep external behavior but use a different parser structure
+
+**Suggested rewrite shape:**
+- `parseVersionString()`
+- `splitReleaseSections()`
+- `toChangelogEntry()`
+- `isVersionNewer()`
+
+#### `src/utils/photon.ts` — similarity: 17.4%, difference: 82.6%, status: DIFFERENT
+
+**Why important:** bigger utility with real logic; more provenance value if rewritten.
+
+**Minimal but meaningful rewrite:**
+- break the wasm fallback flow into separate helpers
+- separate candidate path generation from fs patching
+- separate module caching from load orchestration
+
+**Suggested rewrite shape:**
+- `resolvePhotonWasmCandidates()`
+- `installPhotonReadPatch()`
+- `loadPhotonModuleOnce()`
+- `loadPhoton()` as public wrapper
+
+**Avoid:**
+- only renaming internal variables like `photonModule` / `loadPromise`
+
+#### `src/modes/interactive/components/diff.ts` — similarity: 10.1%, difference: 89.9%, status: DIFFERENT
+
+**Why important:** this is the main P1 code file where structure matters.
+
+**Minimal but meaningful rewrite:**
+- keep the same output contract, but redesign internals
+- separate parsing, pairing, highlighting, and rendering phases
+- use an intermediate row model instead of inline loop-heavy rendering
+
+**Suggested rewrite shape:**
+- `tokenizeDiffLines()`
+- `groupDiffBlocks()`
+- `highlightChangedLinePair()`
+- `renderDiffRows()`
+- `renderDiff()` becomes orchestration only
+
+#### `src/core/export-html/template.html` — similarity: 43.1%, difference: 56.9%, status: DIFFERENT
+
+**Minimal but meaningful rewrite:**
+- change the DOM structure and placeholder ordering
+- rename IDs/classes to your own layout vocabulary
+- reorganize sidebar/header/message sections while keeping runtime compatibility
+
+**Suggested rewrite:**
+- new sidebar markup
+- new toolbar/header structure
+- different search/filter button layout
+- different script/style injection order if safe
+
+#### `src/core/export-html/vendor/highlight.min.js` — similarity: 68.6%, difference: 31.4%, status: DIFFERENT
+#### `src/core/export-html/vendor/marked.min.js` — similarity: 7.5%, difference: 92.5%, status: DIFFERENT
+
+These should **not** be hand-edited.
+
+**Recommended replacement plan:**
+- remove copied versions from repo provenance chain
+- source them directly from the real upstream package or release asset
+- document exact upstream version and license separately
+- wire build/copy scripts to fetch or copy from upstream-installed packages
+
+### P1 rule of thumb
+
+**Not enough:**
+- rename variables
+- change comments
+- change quote style
+- change formatting
+- reorder imports only
+
+**Good enough for a P1 rewrite pass:**
+- new helper decomposition
+- new export/module structure
+- new control-flow shape
+- new intermediate data model
+- same behavior implemented through a different code path
+- vendor assets replaced from direct upstream instead of PI Mono copies
+
+
+
+---
+
+
+
+## P2 — Highest-priority likely-derived runtime files
+
+
+
+These files are not byte-identical, but they share the same path and likely share heavy structural derivation risk. Rewrite these next.
+
+
+
+### P2-A. App/bootstrap/runtime entrypoints
+
+- `src/cli.ts` — similarity: 40.7%, difference: 59.3%, status: DIFFERENT
+
+- `src/config.ts` — similarity: 80.4%, difference: 19.6%, status: DIFFERENT
+
+- `src/index.ts` — similarity: 6.7%, difference: 93.3%, status: DIFFERENT
+
+- `src/main.ts` — similarity: 57.0%, difference: 43.0%, status: DIFFERENT
+
+- `src/migrations.ts` — similarity: 12.4%, difference: 87.6%, status: DIFFERENT
+
+
+
+### P2-B. CLI subsystem
+
+- `src/cli/args.ts` — similarity: 45.2%, difference: 54.8%, status: DIFFERENT
+
+- `src/cli/config-selector.ts` — similarity: 17.1%, difference: 82.9%, status: DIFFERENT
+
+- `src/cli/file-processor.ts` — similarity: 8.9%, difference: 91.1%, status: DIFFERENT
+
+- `src/cli/list-models.ts` — similarity: 13.7%, difference: 86.3%, status: DIFFERENT
+
+- `src/cli/session-picker.ts` — similarity: 8.3%, difference: 91.7%, status: DIFFERENT
+
+
+
+### P2-C. Core session/runtime orchestration
+
+- `src/core/agent-session.ts` — similarity: 75.2%, difference: 24.8%, status: DIFFERENT
+
+- `src/core/auth-storage.ts` — similarity: 42.3%, difference: 57.7%, status: DIFFERENT
+
+- `src/core/bash-executor.ts` — similarity: 6.6%, difference: 93.4%, status: DIFFERENT
+
+- `src/core/diagnostics.ts` — similarity: 11.3%, difference: 88.7%, status: DIFFERENT
+
+- `src/core/event-bus.ts` — similarity: 28.6%, difference: 71.4%, status: DIFFERENT
+
+- `src/core/exec.ts` — similarity: 19.8%, difference: 80.2%, status: DIFFERENT
+
+- `src/core/footer-data-provider.ts` — similarity: 4.2%, difference: 95.8%, status: DIFFERENT
+
+- `src/core/index.ts` — similarity: 3.5%, difference: 96.5%, status: DIFFERENT
+
+- `src/core/keybindings.ts` — similarity: 13.2%, difference: 86.8%, status: DIFFERENT
+
+- `src/core/messages.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/model-registry.ts` — similarity: 64.1%, difference: 35.9%, status: DIFFERENT
+
+- `src/core/model-resolver.ts` — similarity: 7.4%, difference: 92.6%, status: DIFFERENT
+
+- `src/core/package-manager.ts` — similarity: 3.4%, difference: 96.6%, status: DIFFERENT
+
+- `src/core/prompt-templates.ts` — similarity: 99.3%, difference: 0.7%, status: DIFFERENT
+
+- `src/core/resource-loader.ts` — similarity: 84.7%, difference: 15.3%, status: DIFFERENT
+
+- `src/core/sdk.ts` — similarity: 70.9%, difference: 29.1%, status: DIFFERENT
+
+- `src/core/session-manager.ts` — similarity: 93.2%, difference: 6.8%, status: DIFFERENT
+
+- `src/core/settings-manager.ts` — similarity: 61.7%, difference: 38.3%, status: DIFFERENT
+
+- `src/core/skills.ts` — similarity: 82.9%, difference: 17.1%, status: DIFFERENT
+
+- `src/core/system-prompt.ts` — similarity: 53.3%, difference: 46.7%, status: DIFFERENT
+
+- `src/core/timings.ts` — similarity: 32.0%, difference: 68.0%, status: DIFFERENT
+
+
+
+### P2-D. Compaction/runtime helpers
+
+- `src/core/compaction/branch-summarization.ts` — similarity: 98.6%, difference: 1.4%, status: DIFFERENT
+
+- `src/core/compaction/compaction.ts` — similarity: 98.5%, difference: 1.5%, status: DIFFERENT
+
+- `src/core/compaction/utils.ts` — similarity: 92.9%, difference: 7.1%, status: DIFFERENT
+
+
+
+### Why P2 is critical
+
+If these remain structurally PI Mono-derived, then even rewriting surface docs or utilities will not materially change provenance risk.
+
+### Focused P2 plan for the currently requested files
+
+**P2 status update:** the planned rewrite batch for these focused P2 files has now been executed in `indusagi-screth`, and the codebase still builds/tests cleanly afterward.
+
+### Focused P2 post-rewrite recheck (approximate)
+
+**Method note:** the percentages below are an approximate **post-P2** recheck against `@pi-mono/packages/coding-agent/`, using a `git diff --no-index --word-diff=porcelain` unchanged-text ratio. These numbers are meant for provenance prioritization, not as a legal or plagiarism determination.
+
+| File | Post-P2 similarity | Post-P2 difference |
+|---|---:|---:|
+| `src/core/timings.ts` | `15.9%` | `84.1%` |
+| `src/cli.ts` | `11.3%` | `88.7%` |
+| `src/cli/args.ts` | `10.0%` | `90.0%` |
+| `src/config.ts` | `31.0%` | `69.0%` |
+| `src/core/auth-storage.ts` | `26.8%` | `73.2%` |
+| `src/core/system-prompt.ts` | `16.7%` | `83.3%` |
+| `src/core/model-registry.ts` | `23.5%` | `76.5%` |
+| `src/core/prompt-templates.ts` | `30.9%` | `69.1%` |
+| `src/core/skills.ts` | `28.2%` | `71.8%` |
+| `src/core/settings-manager.ts` | `22.8%` | `77.2%` |
+| `src/core/resource-loader.ts` | `34.4%` | `65.6%` |
+| `src/core/sdk.ts` | `30.5%` | `69.5%` |
+| `src/core/session-manager.ts` | `15.2%` | `84.8%` |
+| `src/main.ts` | `25.9%` | `74.1%` |
+| `src/core/agent-session.ts` | `21.0%` | `79.0%` |
+
+### Focused P2 post-rewrite takeaway
+
+This recheck suggests the biggest previously high-similarity P2 files have moved materially away from PI Mono structure in this batch. In particular:
+
+- `src/core/prompt-templates.ts` moved from the earlier near-match risk zone to `30.9%` similarity / `69.1%` difference.
+- `src/core/session-manager.ts` moved from the earlier near-match risk zone to `15.2%` similarity / `84.8%` difference.
+- The remaining focused P2 files in this batch now sit broadly in the `10%`–`34%` similarity range under this approximation.
+
+The files you called out are spread across bootstrap, CLI parsing, runtime orchestration, config, prompt loading, and persistence. The safest plan is to rewrite them in **dependency order**, not by raw similarity alone.
+
+### Focused P2 execution order
+
+#### Wave 1 — low-blast-radius foundation first
+1. `src/core/timings.ts` — similarity: 32.0%, difference: 68.0%, status: DIFFERENT
+2. `src/cli.ts` — similarity: 40.7%, difference: 59.3%, status: DIFFERENT
+3. `src/cli/args.ts` — similarity: 45.2%, difference: 54.8%, status: DIFFERENT
+4. `src/config.ts` — similarity: 80.4%, difference: 19.6%, status: DIFFERENT
+5. `src/core/auth-storage.ts` — similarity: 42.3%, difference: 57.7%, status: DIFFERENT
+6. `src/core/system-prompt.ts` — similarity: 53.3%, difference: 46.7%, status: DIFFERENT
+
+#### Wave 2 — content/model/config loading layer
+7. `src/core/model-registry.ts` — similarity: 64.1%, difference: 35.9%, status: DIFFERENT
+8. `src/core/prompt-templates.ts` — similarity: 99.3%, difference: 0.7%, status: DIFFERENT
+9. `src/core/skills.ts` — similarity: 82.9%, difference: 17.1%, status: DIFFERENT
+10. `src/core/settings-manager.ts` — similarity: 61.7%, difference: 38.3%, status: DIFFERENT
+11. `src/core/resource-loader.ts` — similarity: 84.7%, difference: 15.3%, status: DIFFERENT
+
+#### Wave 3 — runtime/session API layer
+12. `src/core/sdk.ts` — similarity: 70.9%, difference: 29.1%, status: DIFFERENT
+13. `src/core/session-manager.ts` — similarity: 93.2%, difference: 6.8%, status: DIFFERENT
+
+#### Wave 4 — top-level orchestration last
+14. `src/main.ts` — similarity: 57.0%, difference: 43.0%, status: DIFFERENT
+15. `src/core/agent-session.ts` — similarity: 75.2%, difference: 24.8%, status: DIFFERENT
+
+### Focused P2 per-file rewrite plan
+
+#### `src/core/timings.ts` — similarity: 32.0%, difference: 68.0%, status: DIFFERENT
+**Why first:** tiny file, quick provenance win, almost zero integration risk.
+
+**Minimal but meaningful rewrite:**
+- replace current module-level mutable state with a small tracker object
+- add `resetTimings()` helper
+- move formatting into a dedicated printer helper
+- keep `time()` and `printTimings()` public API stable
+
+**Suggested rewrite shape:**
+- `createTimingRecorder()`
+- `recordTiming()`
+- `formatTimingReport()`
+- `printTimings()`
+
+#### `src/cli.ts` — similarity: 40.7%, difference: 59.3%, status: DIFFERENT
+**Why early:** small entrypoint, easy to rewrite fully.
+
+**Minimal but meaningful rewrite:**
+- move MCP log suppression into named helper(s)
+- avoid top-level anonymous patch structure copied from PI Mono style
+- introduce a `bootstrapCli()` wrapper that owns process-title + logger patch + main invocation
+
+**Suggested rewrite shape:**
+- `shouldSuppressMcpLogs()`
+- `installCliLogFilter()`
+- `bootstrapCli()`
+
+#### `src/cli/args.ts` — similarity: 45.2%, difference: 54.8%, status: DIFFERENT
+**Why now:** it affects `main.ts`, but is still isolated enough to rewrite before orchestration files.
+
+**Minimal but meaningful rewrite:**
+- split parser into token readers, validators, and normalizers
+- separate flag alias table from parse loop
+- separate tool parsing / model parsing / extension flag parsing into helpers
+- keep external `Args`, `parseArgs()`, and `printHelp()` API compatible where possible
+
+**Suggested rewrite shape:**
+- `consumeFlagValue()`
+- `parseModeFlag()`
+- `parseToolList()`
+- `parseUnknownExtensionFlag()`
+- `buildHelpSections()`
+
+#### `src/config.ts` — similarity: 80.4%, difference: 19.6%, status: DIFFERENT
+**Why early despite high similarity:** many other files depend on it, so it should be stabilized before deeper rewrites.
+
+**Minimal but meaningful rewrite:**
+- separate package-root detection from exported path helpers
+- centralize env var resolution and tilde expansion
+- group asset path helpers vs user config path helpers vs metadata exports
+- reduce structural resemblance to current PI Mono layout even if behavior stays the same
+
+**Suggested rewrite shape:**
+- `resolvePackageRoot()`
+- `resolveBundledAssetDir()`
+- `expandHomeDir()`
+- `resolveAgentDirFromEnv()`
+
+#### `src/core/auth-storage.ts` — similarity: 42.3%, difference: 57.7%, status: DIFFERENT
+**Why early:** medium complexity, low coupling compared with session runtime.
+
+**Minimal but meaningful rewrite:**
+- split disk I/O, migration, account selection, runtime override logic, and token refresh into separate helpers
+- replace monolithic class flow with clear layers
+- preserve public class API if possible
+
+**Suggested rewrite shape:**
+- `loadAuthFile()`
+- `normalizeAuthData()`
+- `selectAccountId()`
+- `persistAuthData()`
+- `resolveCredentialSource()`
+
+#### `src/core/system-prompt.ts` — similarity: 53.3%, difference: 46.7%, status: DIFFERENT
+**Why early:** many runtime files depend on it, but it is still very rewriteable.
+
+**Minimal but meaningful rewrite:**
+- replace current one-pass assembly with section builders
+- split tool list rendering, guideline rendering, docs path rendering, context rendering, skills rendering
+- keep final `buildSystemPrompt()` export stable
+
+**Suggested rewrite shape:**
+- `buildToolSection()`
+- `buildGuidelinesSection()`
+- `buildDocsSection()`
+- `buildProjectContextSection()`
+- `buildSkillsSection()`
+- `assembleSystemPrompt()`
+
+#### `src/core/model-registry.ts` — similarity: 64.1%, difference: 35.9%, status: DIFFERENT
+**Minimal but meaningful rewrite:**
+- split schema validation, built-in provider loading, custom provider registration, and API key resolution into separate units
+- replace current config-loading order implementation shape with explicit phases
+- keep public registry behavior compatible
+
+**Suggested rewrite shape:**
+- `loadBuiltInModels()`
+- `loadCustomModelsConfig()`
+- `applyProviderOverrides()`
+- `resolveModelApiKey()`
+- `registerExtensionProvider()`
+
+#### `src/core/prompt-templates.ts` — similarity: 99.3%, difference: 0.7%, status: DIFFERENT
+**This is the biggest P2 risk by percentage.**
+
+**Minimal but meaningful rewrite:**
+- replace current parser flow with a clearer staged pipeline
+- separate arg tokenization, placeholder substitution, file loading, and directory discovery
+- preserve the same public function names if needed, but rewrite the internal implementation shape substantially
+
+**Suggested rewrite shape:**
+- `tokenizeTemplateArguments()`
+- `applyTemplateArguments()`
+- `readPromptTemplateFile()`
+- `discoverPromptTemplateFiles()`
+- `mergePromptSources()`
+
+#### `src/core/skills.ts` — similarity: 82.9%, difference: 17.1%, status: DIFFERENT
+**Minimal but meaningful rewrite:**
+- separate validation rules from directory traversal
+- separate skill parsing from diagnostics emission
+- replace recursive loader flow with a discovery + transform + validate pipeline
+
+**Suggested rewrite shape:**
+- `discoverSkillCandidateFiles()`
+- `parseSkillDocument()`
+- `validateSkillFrontmatter()`
+- `buildSkillDiagnostics()`
+
+#### `src/core/settings-manager.ts` — similarity: 61.7%, difference: 38.3%, status: DIFFERENT
+**Minimal but meaningful rewrite:**
+- split file persistence, project/global merge logic, getter/setter helpers, and path normalization
+- avoid one giant stateful class implementation style
+- keep public API compatibility where current runtime depends on it
+
+**Suggested rewrite shape:**
+- `loadSettingsFile()`
+- `mergeSettingsLayers()`
+- `normalizeSettingsPaths()`
+- `createSettingsSnapshot()`
+- `persistSettingsLayer()`
+
+#### `src/core/resource-loader.ts` — similarity: 84.7%, difference: 15.3%, status: DIFFERENT
+**Why after settings/skills/prompts:** it composes those systems.
+
+**Minimal but meaningful rewrite:**
+- replace current class-heavy load orchestration with explicit reload phases
+- separate context file loading, extension loading, hook loading, skill loading, prompt loading, and theme loading into dedicated helpers
+- keep `DefaultResourceLoader` public surface compatible if needed
+
+**Suggested rewrite shape:**
+- `loadProjectContextFiles()`
+- `loadExtensionResources()`
+- `loadHookResources()`
+- `loadSkillResources()`
+- `loadPromptResources()`
+- `loadThemeResources()`
+- `composeResourceSnapshot()`
+
+#### `src/core/sdk.ts` — similarity: 70.9%, difference: 29.1%, status: DIFFERENT
+**Why later:** it wires together many subsystems, so it should be rewritten after those subsystems are stabilized.
+
+**Minimal but meaningful rewrite:**
+- split default creation of auth/model/settings/resource/session into separate factories
+- separate MCP/memory initialization from base agent session construction
+- keep `createAgentSession()` and its return shape compatible if possible
+
+**Suggested rewrite shape:**
+- `createDefaultAuthStorage()`
+- `createDefaultRegistry()`
+- `createDefaultResourceLoader()`
+- `createBaseTools()`
+- `attachOptionalMcpTools()`
+- `attachOptionalMemoryTools()`
+
+#### `src/core/session-manager.ts` — similarity: 93.2%, difference: 6.8%, status: DIFFERENT
+**This is one of the highest-priority P2 rewrites by similarity.**
+
+**Minimal but meaningful rewrite:**
+- keep file format compatibility but redesign internal structure
+- split file I/O, branch traversal, entry indexing, tree reconstruction, and append operations into separate helpers
+- avoid leaving the same method/layout structure even if behavior remains the same
+
+**Suggested rewrite shape:**
+- `readSessionHeader()`
+- `readSessionEntries()`
+- `buildEntryIndex()`
+- `buildSessionTree()`
+- `appendSessionEntry()`
+- `resolveBranchPath()`
+
+#### `src/main.ts` — similarity: 57.0%, difference: 43.0%, status: DIFFERENT
+**Why near the end:** it depends on many of the files above.
+
+**Minimal but meaningful rewrite:**
+- separate CLI command dispatch, model loading, package/resource initialization, session creation, and mode dispatch
+- replace current top-to-bottom orchestration flow with staged command handlers
+
+**Suggested rewrite shape:**
+- `handleUtilityCommands()`
+- `resolveStartupResources()`
+- `createSessionFromArgs()`
+- `runSelectedMode()`
+- `shutdownRuntime()`
+
+#### `src/core/agent-session.ts` — similarity: 75.2%, difference: 24.8%, status: DIFFERENT
+**Save for last in this batch.** It is high-value but very high-blast-radius.
+
+**Minimal but meaningful rewrite:**
+- split this class by concern before deep logic rewrites
+- move tool/runtime rebuild logic, event dispatch, retry logic, compaction logic, and session switching helpers into grouped private helpers or companion modules
+- keep public behavior stable while changing internal architecture substantially
+
+**Suggested rewrite shape:**
+- `buildRuntimeState()`
+- `wireAgentEventHandlers()`
+- `handlePromptFlow()`
+- `handleRetryFlow()`
+- `handleCompactionFlow()`
+- `handleSessionSwitchFlow()`
+
+### Focused P2 rule of thumb
+
+**Not enough:**
+- renaming methods only
+- moving code blocks around without redesigning them
+- changing comments/docstrings only
+- changing type names while leaving control flow the same
+
+**Good enough for a meaningful P2 rewrite pass:**
+- introducing new helper layers
+- changing orchestration phases and boundaries
+- splitting monoliths into loaders/builders/assemblers
+- preserving API behavior while changing implementation structure substantially
+- keeping file-format compatibility without keeping the same internal architecture
+
+### Focused P2 practical note
+
+For this batch, the two biggest hidden-risk files are:
+- `src/core/prompt-templates.ts` — because its similarity is still `99.3%`
+- `src/core/session-manager.ts` — because its similarity is still `93.2%`
+
+So even if `agent-session.ts` is more important functionally, the **highest-priority provenance wins** inside this batch are:
+1. `src/core/prompt-templates.ts`
+2. `src/core/session-manager.ts`
+3. `src/core/resource-loader.ts`
+4. `src/core/skills.ts`
+5. `src/config.ts`
+
+---
+
+
+
+## P3 — Extension / tools / mode system rewrite group
+
+
+
+These are also high-risk because they define the user-visible architecture and behavior.
+
+
+
+### P3-A. Extensions subsystem
+
+- `src/core/extensions/index.ts` — similarity: 87.9%, difference: 12.1%, status: DIFFERENT
+
+- `src/core/extensions/loader.ts` — similarity: 89.2%, difference: 10.8%, status: DIFFERENT
+
+- `src/core/extensions/runner.ts` — similarity: 79.5%, difference: 20.5%, status: DIFFERENT
+
+- `src/core/extensions/types.ts` — similarity: 87.0%, difference: 13.0%, status: DIFFERENT
+
+- `src/core/extensions/wrapper.ts` — similarity: 26.0%, difference: 74.0%, status: DIFFERENT
+
+
+
+### P3-B. Tools subsystem
+
+- `src/core/tools/bash.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/tools/edit-diff.ts` — similarity: 3.7%, difference: 96.3%, status: DIFFERENT
+
+- `src/core/tools/edit.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/tools/find.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/tools/grep.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/tools/index.ts` — similarity: 26.1%, difference: 73.9%, status: DIFFERENT
+
+- `src/core/tools/ls.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/tools/path-utils.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/tools/read.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/tools/truncate.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/core/tools/write.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+
+
+### P3-C. Interactive mode root
+
+- `src/modes/interactive/interactive-mode.ts` — similarity: 69.4%, difference: 30.6%, status: DIFFERENT
+
+- `src/modes/print-mode.ts` — similarity: 22.0%, difference: 78.0%, status: DIFFERENT
+
+- `src/modes/rpc/rpc-client.ts` — similarity: 48.9%, difference: 51.1%, status: DIFFERENT
+
+- `src/modes/rpc/rpc-mode.ts` — similarity: 51.0%, difference: 49.0%, status: DIFFERENT
+
+- `src/modes/rpc/rpc-types.ts` — similarity: 52.4%, difference: 47.6%, status: DIFFERENT
+
+
+
+### P3-D. Theme system
+
+- `src/modes/interactive/theme/theme.ts` — similarity: 83.5%, difference: 16.5%, status: DIFFERENT
+
+- `src/modes/interactive/theme/dark.json` — similarity: 98.8%, difference: 1.2%, status: DIFFERENT
+
+- `src/modes/interactive/theme/light.json` — similarity: 98.8%, difference: 1.2%, status: DIFFERENT
+
+- `src/modes/interactive/theme/theme-schema.json` — similarity: 99.4%, difference: 0.6%, status: DIFFERENT
+
+### Focused P3 plan for the currently requested files
+
+The files you called out span three tightly coupled subsystems:
+- extension registration/runtime contracts
+- RPC mode transport/types
+- interactive mode + theme runtime/assets
+
+Because these drive the visible architecture of the product, the safest plan is to rewrite them in **contract order**: first extension types/runtime contracts, then loaders/runners, then RPC protocol/runtime, then interactive root, then theme runtime/assets.
+
+### Focused P3 execution order
+
+#### Wave 1 — extension contract layer first
+1. `src/core/extensions/types.ts` — similarity: 87.0%, difference: 13.0%, status: DIFFERENT
+2. `src/core/extensions/index.ts` — similarity: 87.9%, difference: 12.1%, status: DIFFERENT
+
+#### Wave 2 — extension runtime/loading layer
+3. `src/core/extensions/loader.ts` — similarity: 89.2%, difference: 10.8%, status: DIFFERENT
+4. `src/core/extensions/runner.ts` — similarity: 79.5%, difference: 20.5%, status: DIFFERENT
+
+#### Wave 3 — RPC protocol and transport layer
+5. `src/modes/rpc/rpc-types.ts` — similarity: 52.4%, difference: 47.6%, status: DIFFERENT
+6. `src/modes/rpc/rpc-client.ts` — similarity: 48.9%, difference: 51.1%, status: DIFFERENT
+7. `src/modes/rpc/rpc-mode.ts` — similarity: 51.0%, difference: 49.0%, status: DIFFERENT
+
+#### Wave 4 — interactive root layer
+8. `src/modes/interactive/interactive-mode.ts` — similarity: 69.4%, difference: 30.6%, status: DIFFERENT
+
+#### Wave 5 — theme runtime and assets last in this subset
+9. `src/modes/interactive/theme/theme.ts` — similarity: 83.5%, difference: 16.5%, status: DIFFERENT
+10. `src/modes/interactive/theme/dark.json` — similarity: 98.8%, difference: 1.2%, status: DIFFERENT
+11. `src/modes/interactive/theme/light.json` — similarity: 98.8%, difference: 1.2%, status: DIFFERENT
+12. `src/modes/interactive/theme/theme-schema.json` — similarity: 99.4%, difference: 0.6%, status: DIFFERENT
+
+### Focused P3 per-file rewrite plan
+
+#### `src/core/extensions/types.ts` — similarity: 87.0%, difference: 13.0%, status: DIFFERENT
+**Why first:** this is the contract surface for the whole extension system. If it stays structurally close, loader/runner rewrites do not help enough.
+
+**Minimal but meaningful rewrite:**
+- split giant type surface into grouped sections with clearer ownership boundaries
+- separate UI contracts, runtime contracts, event contracts, tool contracts, and provider contracts into distinct type clusters
+- replace long flat export flow with structured helper aliases / grouped interfaces while preserving external names where needed
+
+**Suggested rewrite shape:**
+- `ExtensionUiContracts`
+- `ExtensionRuntimeContracts`
+- `ExtensionEventContracts`
+- `ExtensionToolContracts`
+- `ExtensionProviderContracts`
+
+#### `src/core/extensions/index.ts` — similarity: 87.9%, difference: 12.1%, status: DIFFERENT
+**Why early:** barrel files can preserve architecture shape very strongly.
+
+**Minimal but meaningful rewrite:**
+- stop mirroring the original export grouping/layout
+- replace large flat export block with curated grouped exports and explicit public-surface comments
+- restructure the barrel around runtime roles instead of source-file mirroring
+
+**Suggested rewrite shape:**
+- `export type { ... }` grouped by concern
+- `export { ... }` grouped by runtime, wrapping, loading
+- explicit “public API” subsections
+
+#### `src/core/extensions/loader.ts` — similarity: 89.2%, difference: 10.8%, status: DIFFERENT
+**This is one of the highest-priority P3 rewrites.**
+
+**Minimal but meaningful rewrite:**
+- separate path normalization, module resolution, runtime creation, API creation, module loading, and conflict-safe registration into distinct helpers
+- reduce the current long procedural flow into staged loading phases
+- keep Bun/node compatibility behavior, but change implementation boundaries substantially
+
+**Suggested rewrite shape:**
+- `resolveExtensionModulePath()`
+- `createBundledVirtualModules()`
+- `createExtensionApiBridge()`
+- `loadExtensionModule()`
+- `loadExtensionPaths()`
+- `finalizeExtensionLoadResult()`
+
+#### `src/core/extensions/runner.ts` — similarity: 79.5%, difference: 20.5%, status: DIFFERENT
+**Why after loader:** runner depends on the contracts and loaded extension shape.
+
+**Minimal but meaningful rewrite:**
+- separate lifecycle binding, event dispatch, command execution, tool wrapping, UI binding, and shortcut validation into distinct helper layers
+- reduce the single large stateful class feel by moving internal policies into helper functions/private sections
+- keep public runner behavior stable while changing internal orchestration order
+
+**Suggested rewrite shape:**
+- `bindRuntimeActions()`
+- `bindCommandActions()`
+- `emitExtensionLifecycleEvent()`
+- `dispatchExtensionEvent()`
+- `validateShortcutConflicts()`
+- `buildCommandContext()`
+
+#### `src/modes/rpc/rpc-types.ts` — similarity: 52.4%, difference: 47.6%, status: DIFFERENT
+**Why before rpc-client/mode:** this is the protocol contract.
+
+**Minimal but meaningful rewrite:**
+- split command types, response types, event types, extension-ui types, and helper guards into separate sections
+- replace current long narrative-driven structure with protocol-domain grouping
+- preserve wire compatibility while changing type organization substantially
+
+**Suggested rewrite shape:**
+- `RpcCommandTypes`
+- `RpcResponseTypes`
+- `RpcEventTypes`
+- `RpcExtensionUiTypes`
+- `isRpcCommand()` / `isRpcResponse()` helpers
+
+#### `src/modes/rpc/rpc-client.ts` — similarity: 48.9%, difference: 51.1%, status: DIFFERENT
+**Minimal but meaningful rewrite:**
+- separate process spawning, stdout parsing, request tracking, event subscription, timeout handling, and shutdown cleanup into dedicated helpers
+- reduce big sequential client flow into transport/controller pieces
+- keep external client API stable if possible
+
+**Suggested rewrite shape:**
+- `spawnRpcProcess()`
+- `createPendingRequestStore()`
+- `handleRpcStdoutLine()`
+- `handleRpcResponse()`
+- `handleRpcEvent()`
+- `shutdownRpcProcess()`
+
+#### `src/modes/rpc/rpc-mode.ts` — similarity: 51.0%, difference: 49.0%, status: DIFFERENT
+**Minimal but meaningful rewrite:**
+- split stdin command ingestion, JSON parsing, request routing, extension-ui response routing, event emission, and response writing into clear phases
+- preserve protocol behavior while changing server orchestration structure
+
+**Suggested rewrite shape:**
+- `parseRpcInputLine()`
+- `routeRpcCommand()`
+- `handleExtensionUiResponse()`
+- `writeRpcResponse()`
+- `writeRpcEvent()`
+- `runRpcServerLoop()`
+
+#### `src/modes/interactive/interactive-mode.ts` — similarity: 69.4%, difference: 30.6%, status: DIFFERENT
+**Save until after extension/RPC work.** It is the biggest blast-radius file in this subset.
+
+**Minimal but meaningful rewrite:**
+- split layout construction, event subscription, input handling, slash-command handling, picker/dialog launching, and rendering updates into grouped private helpers or companion modules
+- reduce the current giant-controller structure by extracting phase builders first, not by trying to rewrite everything at once
+- preserve user-visible behavior while changing internal module boundaries substantially
+
+**Suggested rewrite shape:**
+- `buildInteractiveLayout()`
+- `subscribeToSessionEvents()`
+- `registerSlashCommands()`
+- `handleEditorSubmission()`
+- `handleInteractiveCommand()`
+- `refreshInteractiveView()`
+
+#### `src/modes/interactive/theme/theme.ts` — similarity: 83.5%, difference: 16.5%, status: DIFFERENT
+**Why later:** runtime depends on theme API, and the asset files should be redesigned together with it.
+
+**Minimal but meaningful rewrite:**
+- separate color resolution, theme loading, theme registry/global state, terminal capability handling, syntax highlighting, and export-color conversion into dedicated helper sections
+- avoid one giant theme-engine file with the same major ordering/layout
+- preserve public theme API names where runtime needs them
+
+**Suggested rewrite shape:**
+- `resolveThemeColorValue()`
+- `detectTerminalColorSupport()`
+- `loadThemeDefinition()`
+- `createThemeRegistry()`
+- `createThemeFormatter()`
+- `createExportColorMap()`
+
+#### `src/modes/interactive/theme/dark.json` — similarity: 98.8%, difference: 1.2%, status: DIFFERENT
+#### `src/modes/interactive/theme/light.json` — similarity: 98.8%, difference: 1.2%, status: DIFFERENT
+#### `src/modes/interactive/theme/theme-schema.json` — similarity: 99.4%, difference: 0.6%, status: DIFFERENT
+**These are the highest-risk exact-shape assets in this requested subset.**
+
+**Minimal but meaningful rewrite:**
+- do **not** just rename a few color variables
+- redesign palette choices, variable naming, ordering, comments/description text, and grouping structure
+- for the schema, rewrite descriptions, property ordering, `$defs` organization, and section layout while preserving validation behavior
+- if needed, add a small internal converter/normalizer so the runtime can consume a restructured schema/theme layout
+
+**Suggested rewrite shape:**
+- `dark.json` / `light.json`: new palette system, new var taxonomy, new grouping order
+- `theme-schema.json`: regroup properties by UI domain, rewrite descriptions, redesign `$defs`
+
+### Focused P3 rule of thumb
+
+**Not enough:**
+- changing only import order
+- renaming a few interfaces/types without changing grouping
+- changing just comments/docstrings in loader/runner/RPC files
+- editing a few color hex values while keeping theme json/schema structure the same
+
+**Good enough for a meaningful P3 rewrite pass:**
+- changing the extension public contract organization
+- changing loader/runner orchestration phases
+- separating RPC protocol contracts from transport/server concerns
+- breaking the interactive root into subsystem builders/dispatchers
+- redesigning theme assets and schema structure, not just palette values
+
+### Focused P3 practical note
+
+For this requested subset, the **highest-priority provenance wins** are:
+1. `src/modes/interactive/theme/theme-schema.json`
+2. `src/modes/interactive/theme/dark.json`
+3. `src/modes/interactive/theme/light.json`
+4. `src/core/extensions/loader.ts`
+5. `src/core/extensions/index.ts`
+6. `src/core/extensions/types.ts`
+7. `src/modes/interactive/theme/theme.ts`
+
+The biggest engineering-risk file in this subset is still:
+- `src/modes/interactive/interactive-mode.ts`
+
+So the safest strategy is: **remove near-exact theme/schema/extension-contract similarity first, then attack the giant interactive root last.**
+
+### Focused P3 execution status update
+
+The planned P3 pass has now been executed across the requested files with structural refactors aimed at provenance separation while keeping runtime behavior intact.
+
+Completed implementation notes:
+- `src/core/extensions/types.ts`
+  - replaced the long event overload block with grouped event/result contract maps plus a generic `on<K extends ExtensionEventName>(...)` surface
+- `src/core/extensions/index.ts`
+  - rebuilt the barrel into runtime-role sections instead of source-file mirroring
+- `src/core/extensions/loader.ts`
+  - split loading into resolver/import/accumulator/finalization helpers
+- `src/core/extensions/runner.ts`
+  - added reusable collection/error helpers and moved repeated lifecycle lookup/error paths behind helper methods
+- `src/modes/rpc/rpc-types.ts`
+  - regrouped the protocol into command/response/UI sections and added runtime type guards
+- `src/modes/rpc/rpc-client.ts`
+  - split spawn/startup/shutdown/request-envelope/response-routing logic into dedicated helpers
+- `src/modes/rpc/rpc-mode.ts`
+  - separated extension UI request emission/response resolution from the line-processing loop
+- `src/modes/interactive/interactive-mode.ts`
+  - extracted startup preparation, header assembly, layout attachment, runtime activation, warning display, and startup-message processing helpers
+- `src/modes/interactive/theme/theme.ts`
+  - added a blueprint-theme normalization path so runtime can consume a more domain-grouped theme source format
+- `src/modes/interactive/theme/dark.json`
+- `src/modes/interactive/theme/light.json`
+- `src/modes/interactive/theme/theme-schema.json`
+  - replaced the near-exact legacy asset shape with a grouped blueprint format (`palette`, `chrome`, `surfaces`, `markdown`, `diff`, `syntax`, `thinking`, `modes`, `exportPalette`) while keeping runtime backward compatibility for legacy themes
+
+Validation after the P3 rewrite pass:
+- `npm run build` ✅
+- `npm test` ✅ (`13` files, `781` tests passed)
+
+### Focused P3 post-rewrite recheck (approximate)
+
+**Method note:** the percentages below are an approximate **post-P3** recheck against `@pi-mono/packages/coding-agent/`, using a `git diff --no-index --word-diff=porcelain` unchanged-text ratio. These numbers are meant for provenance prioritization, not as a legal or plagiarism determination.
+
+| File | Post-P3 similarity | Post-P3 difference |
+|---|---:|---:|
+| `src/core/extensions/types.ts` | `41.1%` | `58.9%` |
+| `src/core/extensions/index.ts` | `35.3%` | `64.7%` |
+| `src/core/extensions/loader.ts` | `57.5%` | `42.5%` |
+| `src/core/extensions/runner.ts` | `36.6%` | `63.4%` |
+| `src/modes/rpc/rpc-types.ts` | `63.8%` | `36.2%` |
+| `src/modes/rpc/rpc-client.ts` | `64.5%` | `35.5%` |
+| `src/modes/rpc/rpc-mode.ts` | `47.3%` | `52.7%` |
+| `src/modes/interactive/interactive-mode.ts` | `56.7%` | `43.3%` |
+| `src/modes/interactive/theme/theme.ts` | `61.6%` | `38.4%` |
+| `src/modes/interactive/theme/dark.json` | `37.6%` | `62.4%` |
+| `src/modes/interactive/theme/light.json` | `35.3%` | `64.7%` |
+| `src/modes/interactive/theme/theme-schema.json` | `45.8%` | `54.2%` |
+
+### Focused P3 post-rewrite takeaway
+
+This recheck suggests the P3 pass materially improved the highest-risk theme asset files and the extension barrel/contracts layer, but the larger runtime-heavy files in this batch still remain above the preferred `<40%` similarity target.
+
+Key takeaways:
+- The near-exact theme asset risk was reduced substantially:
+  - `src/modes/interactive/theme/dark.json` → `37.6%` similarity
+  - `src/modes/interactive/theme/light.json` → `35.3%` similarity
+  - `src/modes/interactive/theme/theme-schema.json` → `45.8%` similarity
+- The extension surface also moved meaningfully:
+  - `src/core/extensions/index.ts` → `35.3%` similarity
+  - `src/core/extensions/runner.ts` → `36.6%` similarity
+  - `src/core/extensions/types.ts` → `41.1%` similarity
+- The most rewrite-resistant files in this batch are still the orchestration-heavy runtimes:
+  - `src/modes/rpc/rpc-client.ts` → `64.5%`
+  - `src/modes/rpc/rpc-types.ts` → `63.8%`
+  - `src/modes/interactive/theme/theme.ts` → `61.6%`
+  - `src/core/extensions/loader.ts` → `57.5%`
+  - `src/modes/interactive/interactive-mode.ts` → `56.7%`
+
+So unlike the focused P2 batch, the focused P3 batch is **not yet all below `40%` similarity** under this approximate method. The next cleanup pass should focus on the still-high orchestration files rather than the already-improved theme JSON assets.
+
+---
+
+
+
+## P4 — Interactive component rewrite batch
+
+
+
+These files are many, but they are important because the TUI layer is highly expressive and likely to preserve PI Mono structure even after edits. Rewrite after P2/P3.
+
+
+
+- `src/modes/interactive/components/armin.ts` — similarity: 99.6%, difference: 0.4%, status: DIFFERENT
+
+- `src/modes/interactive/components/assistant-message.ts` — similarity: 53.0%, difference: 47.0%, status: DIFFERENT
+
+- `src/modes/interactive/components/bash-execution.ts` — similarity: 57.7%, difference: 42.3%, status: DIFFERENT
+
+- `src/modes/interactive/components/bordered-loader.ts` — similarity: 97.7%, difference: 2.3%, status: DIFFERENT
+
+- `src/modes/interactive/components/branch-summary-message.ts` — similarity: 97.4%, difference: 2.6%, status: DIFFERENT
+
+- `src/modes/interactive/components/compaction-summary-message.ts` — similarity: 97.5%, difference: 2.5%, status: DIFFERENT
+
+- `src/modes/interactive/components/config-selector.ts` — similarity: 97.2%, difference: 2.8%, status: DIFFERENT
+
+- `src/modes/interactive/components/countdown-timer.ts` — similarity: 96.1%, difference: 3.9%, status: DIFFERENT
+
+- `src/modes/interactive/components/custom-editor.ts` — similarity: 93.8%, difference: 6.2%, status: DIFFERENT
+
+- `src/modes/interactive/components/custom-message.ts` — similarity: 92.8%, difference: 7.2%, status: DIFFERENT
+
+- `src/modes/interactive/components/dynamic-border.ts` — similarity: 94.1%, difference: 5.9%, status: DIFFERENT
+
+- `src/modes/interactive/components/extension-editor.ts` — similarity: 93.7%, difference: 6.3%, status: DIFFERENT
+
+- `src/modes/interactive/components/extension-input.ts` — similarity: 98.2%, difference: 1.8%, status: DIFFERENT
+
+- `src/modes/interactive/components/extension-selector.ts` — similarity: 98.6%, difference: 1.4%, status: DIFFERENT
+
+- `src/modes/interactive/components/footer.ts` — similarity: 78.5%, difference: 21.5%, status: DIFFERENT
+
+- `src/modes/interactive/components/index.ts` — similarity: 98.4%, difference: 1.6%, status: DIFFERENT
+
+- `src/modes/interactive/components/keybinding-hints.ts` — similarity: 97.7%, difference: 2.3%, status: DIFFERENT
+
+- `src/modes/interactive/components/login-dialog.ts` — similarity: 97.4%, difference: 2.6%, status: DIFFERENT
+
+- `src/modes/interactive/components/model-selector.ts` — similarity: 97.4%, difference: 2.6%, status: DIFFERENT
+
+- `src/modes/interactive/components/oauth-selector.ts` — similarity: 29.3%, difference: 70.7%, status: DIFFERENT
+
+- `src/modes/interactive/components/scoped-models-selector.ts` — similarity: 96.5%, difference: 3.5%, status: DIFFERENT
+
+- `src/modes/interactive/components/session-selector-search.ts` — similarity: 91.6%, difference: 8.4%, status: DIFFERENT
+
+- `src/modes/interactive/components/session-selector.ts` — similarity: 82.5%, difference: 17.5%, status: DIFFERENT
+
+- `src/modes/interactive/components/settings-selector.ts` — similarity: 89.8%, difference: 10.2%, status: DIFFERENT
+
+- `src/modes/interactive/components/show-images-selector.ts` — similarity: 89.6%, difference: 10.4%, status: DIFFERENT
+
+- `src/modes/interactive/components/skill-invocation-message.ts` — similarity: 97.3%, difference: 2.7%, status: DIFFERENT
+
+- `src/modes/interactive/components/theme-selector.ts` — similarity: 92.3%, difference: 7.7%, status: DIFFERENT
+
+- `src/modes/interactive/components/thinking-selector.ts` — similarity: 87.8%, difference: 12.2%, status: DIFFERENT
+
+- `src/modes/interactive/components/tool-execution.ts` — similarity: 66.3%, difference: 33.7%, status: DIFFERENT
+
+- `src/modes/interactive/components/tree-selector.ts` — similarity: 91.0%, difference: 9.0%, status: DIFFERENT
+
+- `src/modes/interactive/components/user-message-selector.ts` — similarity: 99.0%, difference: 1.0%, status: DIFFERENT
+
+- `src/modes/interactive/components/user-message.ts` — similarity: 13.5%, difference: 86.5%, status: DIFFERENT
+
+- `src/modes/interactive/components/visual-truncate.ts` — similarity: 97.0%, difference: 3.0%, status: DIFFERENT
+
+
+
+### Focused P4 plan for the currently requested component files
+
+After reading these component files, the main pattern is clear: a large part of this batch is made of **thin TUI wrappers** around `Container` / `Box` / `SelectList` / `Input` / `Text` with the same border, spacing, hint, and callback structure. That means many of the highest-similarity files here can be pushed down quickly with real structural rewrites, while the larger orchestration-heavy components should be handled later in dependency order.
+
+### Focused P4 execution order
+
+#### Wave 1 — exact-shape tiny components and barrel first
+These are the easiest high-risk wins because they are small and visually near-identical in structure.
+
+1. `src/modes/interactive/components/armin.ts` — similarity: 99.6%, difference: 0.4%, status: DIFFERENT
+2. `src/modes/interactive/components/bordered-loader.ts` — similarity: 97.7%, difference: 2.3%, status: DIFFERENT
+3. `src/modes/interactive/components/branch-summary-message.ts` — similarity: 97.4%, difference: 2.6%, status: DIFFERENT
+4. `src/modes/interactive/components/compaction-summary-message.ts` — similarity: 97.5%, difference: 2.5%, status: DIFFERENT
+5. `src/modes/interactive/components/countdown-timer.ts` — similarity: 96.1%, difference: 3.9%, status: DIFFERENT
+6. `src/modes/interactive/components/dynamic-border.ts` — similarity: 94.1%, difference: 5.9%, status: DIFFERENT
+7. `src/modes/interactive/components/index.ts` — similarity: 98.4%, difference: 1.6%, status: DIFFERENT
+8. `src/modes/interactive/components/keybinding-hints.ts` — similarity: 97.7%, difference: 2.3%, status: DIFFERENT
+9. `src/modes/interactive/components/skill-invocation-message.ts` — similarity: 97.3%, difference: 2.7%, status: DIFFERENT
+10. `src/modes/interactive/components/visual-truncate.ts` — similarity: 97.0%, difference: 3.0%, status: DIFFERENT
+
+#### Wave 2 — selector/dialog wrappers next
+These are still mostly wrapper-shaped, but they each encode product-facing selection flows and should be regrouped by dialog intent rather than current copy-shape.
+
+11. `src/modes/interactive/components/extension-input.ts` — similarity: 98.2%, difference: 1.8%, status: DIFFERENT
+12. `src/modes/interactive/components/extension-selector.ts` — similarity: 98.6%, difference: 1.4%, status: DIFFERENT
+13. `src/modes/interactive/components/login-dialog.ts` — similarity: 97.4%, difference: 2.6%, status: DIFFERENT
+14. `src/modes/interactive/components/model-selector.ts` — similarity: 97.4%, difference: 2.6%, status: DIFFERENT
+15. `src/modes/interactive/components/scoped-models-selector.ts` — similarity: 96.5%, difference: 3.5%, status: DIFFERENT
+16. `src/modes/interactive/components/show-images-selector.ts` — similarity: 89.6%, difference: 10.4%, status: DIFFERENT
+17. `src/modes/interactive/components/theme-selector.ts` — similarity: 92.3%, difference: 7.7%, status: DIFFERENT
+18. `src/modes/interactive/components/thinking-selector.ts` — similarity: 87.8%, difference: 12.2%, status: DIFFERENT
+
+#### Wave 3 — editor/custom-render wrapper layer
+These sit between the TUI primitives and interactive-mode orchestration and should be rethought as reusable surface components, not file-by-file copies.
+
+19. `src/modes/interactive/components/custom-editor.ts` — similarity: 93.8%, difference: 6.2%, status: DIFFERENT
+20. `src/modes/interactive/components/custom-message.ts` — similarity: 92.8%, difference: 7.2%, status: DIFFERENT
+21. `src/modes/interactive/components/extension-editor.ts` — similarity: 93.7%, difference: 6.3%, status: DIFFERENT
+22. `src/modes/interactive/components/user-message-selector.ts` — similarity: 99.0%, difference: 1.0%, status: DIFFERENT
+
+#### Wave 4 — session/settings/navigation selectors
+These are larger stateful list controllers with filtering, grouping, selection persistence, and callback orchestration.
+
+23. `src/modes/interactive/components/config-selector.ts` — similarity: 97.2%, difference: 2.8%, status: DIFFERENT
+24. `src/modes/interactive/components/session-selector-search.ts` — similarity: 91.6%, difference: 8.4%, status: DIFFERENT
+25. `src/modes/interactive/components/session-selector.ts` — similarity: 82.5%, difference: 17.5%, status: DIFFERENT
+26. `src/modes/interactive/components/settings-selector.ts` — similarity: 89.8%, difference: 10.2%, status: DIFFERENT
+27. `src/modes/interactive/components/tree-selector.ts` — similarity: 91.0%, difference: 9.0%, status: DIFFERENT
+28. `src/modes/interactive/components/footer.ts` — similarity: 78.5%, difference: 21.5%, status: DIFFERENT
+
+#### Wave 5 — message/execution renderers last in this subset
+These have the biggest rendering blast radius because they encode live output behavior, markdown flow, async conversion, diff/image handling, and chat display semantics.
+
+29. `src/modes/interactive/components/assistant-message.ts` — similarity: 53.0%, difference: 47.0%, status: DIFFERENT
+30. `src/modes/interactive/components/bash-execution.ts` — similarity: 57.7%, difference: 42.3%, status: DIFFERENT
+31. `src/modes/interactive/components/tool-execution.ts` — similarity: 66.3%, difference: 33.7%, status: DIFFERENT
+
+#### Wave 6 — already materially separated, keep for the end or skip in a first P4 pass
+These are in this batch, but they are not the first provenance priority because their approximate similarity is already much lower than the rest.
+
+32. `src/modes/interactive/components/oauth-selector.ts` — similarity: 29.3%, difference: 70.7%, status: DIFFERENT
+33. `src/modes/interactive/components/user-message.ts` — similarity: 13.5%, difference: 86.5%, status: DIFFERENT
+
+### Focused P4 per-cluster rewrite plan
+
+#### P4-A. Tiny exact-shape display helpers
+Files:
+- `armin.ts`
+- `bordered-loader.ts`
+- `branch-summary-message.ts`
+- `compaction-summary-message.ts`
+- `countdown-timer.ts`
+- `dynamic-border.ts`
+- `index.ts`
+- `keybinding-hints.ts`
+- `skill-invocation-message.ts`
+- `visual-truncate.ts`
+
+**What the code reading shows:** these are mostly direct helper/display wrappers with very concentrated logic, which makes them high-risk because structure and ordering are almost fully preserved.
+
+**Minimal but meaningful rewrite:**
+- replace straight-line constructor/build patterns with helper factories
+- regroup formatting logic so labels, border creation, truncation, and timers are not laid out the same way
+- for `index.ts`, stop doing a straight export mirror of the file tree and group by component purpose
+- for `armin.ts`, keep the easter-egg behavior but rewrite the animation/effect pipeline shape entirely
+
+**Suggested rewrite shape:**
+- `createBorderLine()` / `createHintText()` helpers
+- `buildCollapsedSummaryBody()` / `buildExpandedSummaryBody()` helpers
+- `createCountdownLifecycle()` instead of inline timer constructor flow
+- `renderVisualTail()` / `measureVisualLines()` split for truncate helper
+- `components/index.ts` grouped by `dialogs`, `selectors`, `messages`, `execution`, `support`
+
+#### P4-B. Selector and dialog wrappers
+Files:
+- `extension-input.ts`
+- `extension-selector.ts`
+- `login-dialog.ts`
+- `model-selector.ts`
+- `scoped-models-selector.ts`
+- `show-images-selector.ts`
+- `theme-selector.ts`
+- `thinking-selector.ts`
+
+**What the code reading shows:** most of these share the same structure: top border → title/hint → input/list → footer hints → bottom border. That repeated frame is exactly what should change.
+
+**Minimal but meaningful rewrite:**
+- introduce reusable dialog frame builders so each file no longer hand-builds the same border/spacer/title shell
+- separate item-source preparation from list rendering from keyboard handling
+- for model/scoped-model selectors, separate filtering/sorting/persistence logic from view rendering
+- for login dialog, split auth URL display, prompt collection, waiting state, and cancellation into dedicated internal phases
+
+**Suggested rewrite shape:**
+- `createDialogFrame()`
+- `buildSelectorItems()`
+- `renderSelectionList()`
+- `handleSelectorNavigation()`
+- `buildAuthPromptState()` / `buildAuthWaitingState()`
+- `createModelFilterState()` / `createScopedModelOrderState()`
+
+#### P4-C. Editor/custom-render wrapper layer
+Files:
+- `custom-editor.ts`
+- `custom-message.ts`
+- `extension-editor.ts`
+- `user-message-selector.ts`
+
+**What the code reading shows:** these are wrapper adapters between TUI primitives and app-specific behavior. They should become reusable app surfaces, not one-off copied wrappers.
+
+**Minimal but meaningful rewrite:**
+- split action routing from editor inheritance in `custom-editor.ts`
+- split fallback renderer logic from custom-renderer dispatch in `custom-message.ts`
+- split external-editor integration from layout composition in `extension-editor.ts`
+- split message-list navigation, item formatting, and selection callbacks in `user-message-selector.ts`
+
+**Suggested rewrite shape:**
+- `createEditorActionRouter()`
+- `renderDefaultCustomMessage()` / `renderCustomMessageWithExtension()`
+- `launchExternalEditorSession()` / `restoreEditorSession()`
+- `buildUserMessageRows()` / `renderUserMessageSelection()`
+
+#### P4-D. Session/settings/navigation selectors
+Files:
+- `config-selector.ts`
+- `session-selector-search.ts`
+- `session-selector.ts`
+- `settings-selector.ts`
+- `tree-selector.ts`
+- `footer.ts`
+
+**What the code reading shows:** these are not tiny wrapper files. They are mini state machines with grouping/filtering/rendering rules. They need real decomposition, not cosmetic changes.
+
+**Minimal but meaningful rewrite:**
+- separate search parsing/scoring from session list rendering
+- separate tree flattening, filtering, gutter/connector calculation, and row rendering in `tree-selector.ts`
+- separate resource grouping/path formatting/toggling from the UI shell in `config-selector.ts`
+- separate footer stat aggregation, width allocation, and extension-status rendering in `footer.ts`
+- separate settings item schema construction from submenu rendering in `settings-selector.ts`
+
+**Suggested rewrite shape:**
+- `buildResourceGroups()` / `flattenResourceEntries()` / `toggleResourceSetting()`
+- `parseSessionQuery()` / `scoreSessionResult()` / `renderSessionRows()`
+- `buildTreeRows()` / `filterTreeRows()` / `renderTreeRow()`
+- `collectFooterUsageStats()` / `layoutFooterColumns()` / `renderFooterStatuses()`
+- `buildSettingsItems()` / `createSettingsSubmenu()`
+
+#### P4-E. High-blast-radius execution/message renderers
+Files:
+- `assistant-message.ts`
+- `bash-execution.ts`
+- `tool-execution.ts`
+
+**What the code reading shows:** these are the core message/tool visualization surfaces. They should be rewritten last because they touch the user-visible flow most deeply.
+
+**Minimal but meaningful rewrite:**
+- split pending/result/render state from layout containers
+- separate built-in tool formatters from media/diff/preview handling in `tool-execution.ts`
+- separate output buffering, truncation, and status rendering in `bash-execution.ts`
+- separate content block extraction from thinking/error rendering in `assistant-message.ts`
+
+**Suggested rewrite shape:**
+- `extractAssistantRenderableBlocks()` / `renderAssistantThinking()` / `renderAssistantErrorState()`
+- `appendBashChunks()` / `buildBashPreview()` / `renderBashStatus()`
+- `resolveToolRenderer()` / `formatBuiltInToolResult()` / `renderToolMedia()` / `renderToolPreview()`
+
+### Focused P4-A execution status update
+
+The first P4 execution pass has now been started and the **Wave 1 / P4-A tiny-helper batch** has been structurally rewritten in `indusagi-screth`:
+
+- `src/modes/interactive/components/armin.ts`
+- `src/modes/interactive/components/bordered-loader.ts`
+- `src/modes/interactive/components/branch-summary-message.ts`
+- `src/modes/interactive/components/compaction-summary-message.ts`
+- `src/modes/interactive/components/countdown-timer.ts`
+- `src/modes/interactive/components/dynamic-border.ts`
+- `src/modes/interactive/components/index.ts`
+- `src/modes/interactive/components/keybinding-hints.ts`
+- `src/modes/interactive/components/skill-invocation-message.ts`
+- `src/modes/interactive/components/visual-truncate.ts`
+
+This pass changed the internal implementation shape toward helper/factory/state-based structures rather than the previous straight-line wrapper layout. In particular:
+
+- border rendering now goes through explicit border-line helpers and cached painters
+- keybinding hint formatting now flows through reusable key-resolution and hint-building helpers
+- visual truncation now separates measurement from tail rendering
+- countdown timing now uses a dedicated lifecycle factory instead of inline interval ownership
+- the branch / compaction / skill summary components now build collapsed vs expanded bodies through dedicated helper functions
+- the loader wrapper now uses an explicit loader-binding abstraction rather than inline branching in the constructor
+- the component barrel now groups exports by UI role instead of mirroring the file list only
+- `armin.ts` now uses an effect-state animation pipeline instead of the prior monolithic per-method state object pattern
+
+Validation after the P4-A rewrite batch:
+
+- `cd /Users/varunisrani/indusagi-ts/indusagi-screth && npm run build` ✅
+- `cd /Users/varunisrani/indusagi-ts/indusagi-screth && npm test` ✅
+- `13` test files passed
+- `781` tests passed
+
+A fresh similarity recheck for this P4-A subset still needs to be run after the next rewrite wave or on request.
+
+### Focused remaining P4 execution status update
+
+The next P4 waves have now also been executed in `indusagi-screth`.
+
+#### P4-B selector/dialog wrappers rewritten
+- `src/modes/interactive/components/extension-input.ts`
+- `src/modes/interactive/components/extension-selector.ts`
+- `src/modes/interactive/components/login-dialog.ts`
+- `src/modes/interactive/components/model-selector.ts`
+- `src/modes/interactive/components/scoped-models-selector.ts`
+- `src/modes/interactive/components/show-images-selector.ts`
+- `src/modes/interactive/components/theme-selector.ts`
+- `src/modes/interactive/components/thinking-selector.ts`
+
+This wave moved repeated selector/dialog shells toward helper-driven construction such as timed title bindings, item builders, dialog-state helpers, and explicit list/render routing.
+
+#### P4-C editor/custom-render wrapper layer rewritten
+- `src/modes/interactive/components/custom-editor.ts`
+- `src/modes/interactive/components/custom-message.ts`
+- `src/modes/interactive/components/extension-editor.ts`
+- `src/modes/interactive/components/user-message-selector.ts`
+
+This wave split editor action routing, fallback-vs-extension message rendering, external-editor session handling, and user-message row construction into separate helper layers.
+
+#### P4-D session/settings/navigation selector batch rewritten
+- `src/modes/interactive/components/config-selector.ts`
+- `src/modes/interactive/components/session-selector-search.ts`
+- `src/modes/interactive/components/session-selector.ts`
+- `src/modes/interactive/components/settings-selector.ts`
+- `src/modes/interactive/components/tree-selector.ts`
+- `src/modes/interactive/components/footer.ts`
+
+This wave added deeper decomposition around resource flattening/filter sets, session row rendering, tree prefix/child prioritization helpers, settings item builders/dispatchers, and footer aggregation/layout helpers.
+
+#### P4-E execution/message renderer batch rewritten
+- `src/modes/interactive/components/assistant-message.ts`
+- `src/modes/interactive/components/bash-execution.ts`
+- `src/modes/interactive/components/tool-execution.ts`
+
+This wave split assistant block extraction, bash output buffering/preview/status formatting, and tool rendering state/image/background resolution into more explicit helper phases instead of keeping the earlier flatter renderer structure.
+
+Validation after the broader remaining P4 rewrite pass:
+
+- `cd /Users/varunisrani/indusagi-ts/indusagi-screth && npm run build` ✅
+- `cd /Users/varunisrani/indusagi-ts/indusagi-screth && npm test` ✅
+- `13` test files passed
+- `781` tests passed
+
+A fresh post-P4 similarity recheck has now been run using the same **approximate** `git diff --no-index --word-diff=porcelain` unchanged-text ratio method used for the earlier focused rechecks.
+
+### Focused P4 post-rewrite recheck (approximate)
+
+- `src/modes/interactive/components/armin.ts` → `10.8%` similarity / `89.2%` difference
+- `src/modes/interactive/components/assistant-message.ts` → `15.0%` / `85.0%`
+- `src/modes/interactive/components/bash-execution.ts` → `18.4%` / `81.6%`
+- `src/modes/interactive/components/bordered-loader.ts` → `16.8%` / `83.2%`
+- `src/modes/interactive/components/branch-summary-message.ts` → `32.9%` / `67.1%`
+- `src/modes/interactive/components/compaction-summary-message.ts` → `28.4%` / `71.6%`
+- `src/modes/interactive/components/config-selector.ts` → `36.5%` / `63.5%`
+- `src/modes/interactive/components/countdown-timer.ts` → `10.3%` / `89.7%`
+- `src/modes/interactive/components/custom-editor.ts` → `27.2%` / `72.8%`
+- `src/modes/interactive/components/custom-message.ts` → `23.3%` / `76.7%`
+- `src/modes/interactive/components/dynamic-border.ts` → `22.2%` / `77.8%`
+- `src/modes/interactive/components/extension-editor.ts` → `28.9%` / `71.1%`
+- `src/modes/interactive/components/extension-input.ts` → `15.8%` / `84.2%`
+- `src/modes/interactive/components/extension-selector.ts` → `26.5%` / `73.5%`
+- `src/modes/interactive/components/footer.ts` → `7.7%` / `92.3%`
+- `src/modes/interactive/components/index.ts` → `36.5%` / `63.5%`
+- `src/modes/interactive/components/keybinding-hints.ts` → `31.3%` / `68.7%`
+- `src/modes/interactive/components/login-dialog.ts` → `16.4%` / `83.6%`
+- `src/modes/interactive/components/model-selector.ts` → `23.9%` / `76.1%`
+- `src/modes/interactive/components/oauth-selector.ts` → `13.0%` / `87.0%`
+- `src/modes/interactive/components/scoped-models-selector.ts` → `37.8%` / `62.2%`
+- `src/modes/interactive/components/session-selector-search.ts` → `56.8%` / `43.2%`
+- `src/modes/interactive/components/session-selector.ts` → `42.0%` / `58.0%`
+- `src/modes/interactive/components/settings-selector.ts` → `15.2%` / `84.8%`
+- `src/modes/interactive/components/show-images-selector.ts` → `20.8%` / `79.2%`
+- `src/modes/interactive/components/skill-invocation-message.ts` → `30.4%` / `69.6%`
+- `src/modes/interactive/components/theme-selector.ts` → `22.0%` / `78.0%`
+- `src/modes/interactive/components/thinking-selector.ts` → `26.9%` / `73.1%`
+- `src/modes/interactive/components/tool-execution.ts` → `20.2%` / `79.8%`
+- `src/modes/interactive/components/tree-selector.ts` → `35.9%` / `64.1%`
+- `src/modes/interactive/components/user-message-selector.ts` → `25.7%` / `74.3%`
+- `src/modes/interactive/components/user-message.ts` → `3.0%` / `97.0%`
+- `src/modes/interactive/components/visual-truncate.ts` → `16.0%` / `84.0%`
+
+### Focused P4 post-rewrite takeaway
+
+The P4 batch is now **materially lower-risk than before**. Under this approximate recheck, **31 of 33 files are now below `40%` similarity**.
+
+The only remaining files above the preferred `<40%` target in this focused P4 batch are:
+- `src/modes/interactive/components/session-selector-search.ts` → `56.8%`
+- `src/modes/interactive/components/session-selector.ts` → `42.0%`
+
+So the next P4-specific cleanup, if desired, should focus on the session search/sorting UI and the larger session-selector orchestration file, not the already-reduced wrapper/message/helper files.
+
+### Focused P4 rule of thumb
+
+**Not enough:**
+- renaming classes only
+- moving a few methods around without changing the shell/layout pattern
+- changing only strings, labels, or comments
+- changing a few colors but keeping the same border/title/list constructor sequence
+- changing barrel export order only
+
+**Good enough for a meaningful P4 rewrite pass:**
+- replacing repeated border-dialog shells with a new component frame architecture
+- moving filtering/search/order logic out of UI containers into dedicated helpers
+- replacing flat render/update methods with state/build/render phases
+- regrouping component exports by UI role rather than mirroring file names
+- redesigning tiny high-similarity files so their control flow is no longer line-for-line recognizable
+
+### Focused P4 practical priority note
+
+For this requested subset, the **highest-priority provenance wins** are:
+1. `src/modes/interactive/components/armin.ts`
+2. `src/modes/interactive/components/user-message-selector.ts`
+3. `src/modes/interactive/components/extension-selector.ts`
+4. `src/modes/interactive/components/extension-input.ts`
+5. `src/modes/interactive/components/index.ts`
+6. `src/modes/interactive/components/bordered-loader.ts`
+7. `src/modes/interactive/components/login-dialog.ts`
+8. `src/modes/interactive/components/model-selector.ts`
+9. `src/modes/interactive/components/config-selector.ts`
+10. `src/modes/interactive/components/tree-selector.ts`
+
+The biggest engineering-risk files in this subset are:
+- `src/modes/interactive/components/tool-execution.ts`
+- `src/modes/interactive/components/tree-selector.ts`
+- `src/modes/interactive/components/session-selector.ts`
+- `src/modes/interactive/components/config-selector.ts`
+- `src/modes/interactive/components/footer.ts`
+
+So the safest strategy is:
+- **first remove the near-exact tiny/wrapper component shapes**
+- **then decompose selector/state-machine components**
+- **then rewrite the tool/message rendering cores last**
+
+### Note
+
+You can keep the same product behavior, but the implementation should be independently rethought and rewritten.
+
+
+
+---
+
+
+
+## P5 — Lower-priority but still should change
+
+
+
+These files are less central than the runtime architecture, but they still overlap and should be rewritten/cleaned if the goal is a strong no-PI-Mono claim.
+
+
+
+### P5-A. Utility/support files
+
+- `src/utils/clipboard-image.ts` — similarity: 63.0%, difference: 37.0%, status: DIFFERENT
+
+- `src/utils/clipboard.ts` — similarity: 46.3%, difference: 53.7%, status: DIFFERENT
+
+- `src/utils/frontmatter.ts` — similarity: 0.0%, difference: 100.0%, status: DIFFERENT
+
+- `src/utils/git.ts` — similarity: 2.0%, difference: 98.0%, status: DIFFERENT
+
+- `src/utils/image-convert.ts` — similarity: 93.7%, difference: 6.3%, status: DIFFERENT
+
+- `src/utils/image-resize.ts` — similarity: 98.1%, difference: 1.9%, status: DIFFERENT
+
+- `src/utils/shell.ts` — similarity: 91.1%, difference: 8.9%, status: DIFFERENT
+
+- `src/utils/tools-manager.ts` — similarity: 80.4%, difference: 19.6%, status: DIFFERENT
+
+
+
+### P5-B. Package/build/metadata files
+
+- `package.json`
+
+- `README.md`
+
+- `CHANGELOG.md`
+
+- `tsconfig.build.json`
+
+- `vitest.config.ts`
+
+- `examples/README.md`
+
+
+
+These may not be the biggest copyright risk, but they should be cleaned to remove lineage clues and copied structure if you want a consistent provenance story.
+
+### Focused P5 execution status update
+
+A focused P5 cleanup pass has now been executed for the higher-similarity utility and metadata files.
+
+#### P5-A utility/support files rewritten in this pass
+- `src/utils/clipboard-image.ts`
+- `src/utils/clipboard.ts`
+- `src/utils/image-convert.ts`
+- `src/utils/image-resize.ts`
+- `src/utils/shell.ts`
+- `src/utils/tools-manager.ts`
+
+These were restructured around new helper layers rather than the previous flatter layouts. In particular:
+- clipboard image handling now flows through normalized mime ranking, backend-specific readers, and shared command wrappers
+- clipboard text writing now uses backend writer selection rather than a single linear per-platform branch
+- image conversion/resizing now uses decode/encode/budget/variant helper phases
+- shell resolution now uses explicit configured/windows/posix resolver paths plus isolated process-tree and output-sanitization helpers
+- bundled tool installation now uses a tool registry with separate release lookup, download, extraction, and install phases
+
+#### P5-A files already low-similarity and not priority-edited in this pass
+- `src/utils/frontmatter.ts`
+- `src/utils/git.ts`
+
+Those were already recorded in the guide as materially separated, so they were not the first P5 rewrite priority.
+
+#### P5-B package/build/metadata files cleaned in this pass
+- `package.json`
+- `README.md`
+- `CHANGELOG.md`
+- `tsconfig.build.json`
+- `vitest.config.ts`
+- `examples/README.md`
+
+This pass removed direct reference-style wording, simplified lineage-exposing phrasing, and reorganized the package/build metadata into more project-specific descriptions and layout.
+
+Validation after the P5 rewrite batch:
+
+- `cd /Users/varunisrani/indusagi-ts/indusagi-screth && npm run build` ✅
+- `cd /Users/varunisrani/indusagi-ts/indusagi-screth && npm test` ✅
+- `13` test files passed
+- `781` tests passed
+
+A fresh post-P5 similarity recheck has now been run using the same **approximate** `git diff --no-index --word-diff=porcelain` unchanged-text ratio method used elsewhere in this guide.
+
+### Focused P5 post-rewrite recheck (approximate)
+
+#### P5-A utility/support files
+- `src/utils/clipboard-image.ts` → `15.6%` similarity / `84.4%` difference
+- `src/utils/clipboard.ts` → `6.7%` / `93.3%`
+- `src/utils/frontmatter.ts` → `2.0%` / `98.0%`
+- `src/utils/git.ts` → `1.0%` / `99.0%`
+- `src/utils/image-convert.ts` → `13.3%` / `86.7%`
+- `src/utils/image-resize.ts` → `15.4%` / `84.6%`
+- `src/utils/shell.ts` → `12.3%` / `87.7%`
+- `src/utils/tools-manager.ts` → `18.4%` / `81.6%`
+
+#### P5-B package/build/metadata files
+- `package.json` → `18.8%` similarity / `81.2%` difference
+- `README.md` → `0.2%` / `99.8%`
+- `CHANGELOG.md` → `0.0%` / `100.0%`
+- `tsconfig.build.json` → `12.0%` / `88.0%`
+- `vitest.config.ts` → `12.7%` / `87.3%`
+- `examples/README.md` → `2.3%` / `97.7%`
+
+### Focused P5 post-rewrite takeaway
+
+The focused P5 batch is now in a strong state under this approximate method: **all 14 rechecked P5 files are below `40%` similarity**, and most are far below that threshold.
+
+That means the remaining provenance priority is no longer the P5 utility/metadata layer. The next higher-value work is to keep focusing on any still-high orchestration files left from earlier batches.
+
+### Targeted recheck for the last previously >=50% files (approximate)
+
+After the follow-up rewrite pass on the final high-similarity holdouts, the previously remaining `>=50%` files now recheck as:
+
+- `src/core/event-bus.ts` → `26.8%` similarity / `73.2%` difference
+- `src/core/compaction/compaction.ts` → `44.0%` / `56.0%`
+- `src/core/compaction/branch-summarization.ts` → `44.2%` / `55.8%`
+- `src/modes/interactive/components/session-selector-search.ts` → `19.5%` / `80.5%`
+- `src/core/keybindings.ts` → `19.2%` / `80.8%`
+- `src/core/compaction/utils.ts` → `16.3%` / `83.7%`
+- `src/core/export-html/tool-renderer.ts` → `16.6%` / `83.4%`
+- `src/migrations.ts` → `21.5%` / `78.5%`
+
+### Targeted holdout takeaway
+
+Under this approximate recheck, **all 8 of the files that were previously still at or above `50%` similarity are now below `50%`**.
+
+That materially changes the repository-wide overlap picture: there are no longer any previously identified shared-path holdout files from this batch still above the user’s `<50%` comfort threshold.
+
+---
+
+
+
+# Files that are target-only right now
+
+
+
+These files exist in `indusagi-screth` but not in PI Mono, so they are **not immediate PI Mono overlap evidence** based on path comparison alone:
+
+
+
+- `src/core/discover-packages.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/subagents.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/task-session-manager.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/todo-store.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/tools/bg-process.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/tools/memory.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/tools/registry.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/tools/task.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/tools/todo.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/tools/webfetch.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/core/tools/websearch.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/modes/shared.ts` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/observability/**` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- `src/utils/__tests__/**` — similarity: 0.0%, difference: 100.0%, status: MISSING_IN_ONE
+
+- many additional target-only docs/hooks/skills
+
+
+
+These still deserve provenance review, but they are **not first priority** for PI Mono separation.
+
+
+
+---
+
+
+
+# Recommended execution order
+
+
+
+## Phase 1 — Immediate risk removal
+
+1. Rewrite all P1 exact-match files.
+
+2. Replace vendor assets with direct upstream-managed assets or regenerate them.
+
+3. Add a temporary provenance tracker document.
+
+
+
+## Phase 2 — Core independence
+
+4. Rewrite P2 runtime/core files.
+
+5. Rewrite P3 extensions/tools/modes roots.
+
+
+
+## Phase 3 — UI independence
+
+6. Rewrite P4 interactive components.
+
+7. Rewrite P5 utilities and metadata.
+
+
+
+## Phase 4 — Final cleanup
+
+8. Re-run repo comparison against PI Mono.
+
+9. Confirm there are no exact-match code files left.
+
+10. Review remaining same-path files manually for structural derivation.
+
+11. Update license/NOTICE/readme language accordingly.
+
+
+
+---
+
+
+
+# Definition of done
+
+
+
+You can make a much stronger statement like _"`indusagi-screth` is no longer PI Mono-derived"_ only when all of the following are true:
+
+
+
+- [ ] No exact identical PI Mono code files remain.
+
+- [ ] No copied PI Mono vendor assets remain unless separately sourced and attributed.
+
+- [ ] High-risk core files have been independently rewritten.
+
+- [ ] Interactive mode and component system have been independently rewritten.
+
+- [ ] Utility/support files with overlap have been independently rewritten.
+
+- [ ] Docs/package metadata no longer reflect copied structure.
+
+- [ ] A fresh comparison shows either:
+
+  - no same-path exact matches, or
+
+  - only harmless/generated/upstream-third-party artifacts.
+
+
+
+---
+
+
+
+# Practical recommendation
+
+
+
+If time is limited, start with this minimal high-value shortlist:
+
+
+
+## Minimal shortlist to change first
+
+1. `src/core/agent-session.ts` — similarity: 75.2%, difference: 24.8%, status: DIFFERENT
+
+2. `src/core/session-manager.ts` — similarity: 93.2%, difference: 6.8%, status: DIFFERENT
+
+3. `src/core/sdk.ts` — similarity: 70.9%, difference: 29.1%, status: DIFFERENT
+
+4. `src/core/settings-manager.ts` — similarity: 61.7%, difference: 38.3%, status: DIFFERENT
+
+5. `src/core/resource-loader.ts` — similarity: 84.7%, difference: 15.3%, status: DIFFERENT
+
+6. `src/core/extensions/types.ts` — similarity: 87.0%, difference: 13.0%, status: DIFFERENT
+
+7. `src/core/extensions/runner.ts` — similarity: 79.5%, difference: 20.5%, status: DIFFERENT
+
+8. `src/core/tools/index.ts` — similarity: 26.1%, difference: 73.9%, status: DIFFERENT
+
+9. `src/modes/interactive/interactive-mode.ts` — similarity: 69.4%, difference: 30.6%, status: DIFFERENT
+
+10. `src/modes/interactive/theme/theme.ts` — similarity: 83.5%, difference: 16.5%, status: DIFFERENT
+
+11. `src/modes/rpc/rpc-client.ts` — similarity: 48.9%, difference: 51.1%, status: DIFFERENT
+
+12. `src/modes/rpc/rpc-mode.ts` — similarity: 51.0%, difference: 49.0%, status: DIFFERENT
+
+13. `src/modes/rpc/rpc-types.ts` — similarity: 52.4%, difference: 47.6%, status: DIFFERENT
+
+14. `src/modes/print-mode.ts` — similarity: 22.0%, difference: 78.0%, status: DIFFERENT
+
+15. all P1 exact-match files
+
+
+
+If you want the strongest provenance cleanup, use the full priority list above, not just the minimal shortlist.
+
+
+
+## Appendix A — Approximate Per-File Similarity / Difference Percentages
+
+
+
+These percentages are **approximate line-based estimates**, not legal determinations.
+
+
+
+Method used:
+
+- compare the PI Mono file and the `indusagi-screth` file at the same relative path
+
+- use line counts plus `git diff --no-index --numstat` style added/deleted line counts
+
+- **Similarity %** ≈ unchanged-line ratio
+
+- **Difference %** ≈ changed-line ratio
+
+
+
+Interpretation:
+
+- `100% similarity` = exact file match
+
+- high similarity means a stronger rewrite priority
+
+- low similarity does **not** automatically mean legally clean; structure/derivation risk may still exist
+
+
+
+| File | Status | PI lines | Indus lines | Similarity | Difference |
+
+|---|---:|---:|---:|---:|---:|
+
+| `src/core/compaction/index.ts` | DIFFERENT | 7 | 98 | 5.7% | 94.3% |
+
+| `src/modes/index.ts` | DIFFERENT | 9 | 38 | 12.8% | 87.2% |
+
+| `src/modes/interactive/components/diff.ts` | DIFFERENT | 147 | 171 | 10.1% | 89.9% |
+
+| `src/utils/changelog.ts` | DIFFERENT | 99 | 103 | 25.7% | 74.3% |
+
+| `src/utils/mime.ts` | DIFFERENT | 30 | 53 | 16.9% | 83.1% |
+
+| `src/utils/photon.ts` | DIFFERENT | 139 | 159 | 17.4% | 82.6% |
+
+| `src/utils/sleep.ts` | DIFFERENT | 18 | 47 | 18.5% | 81.5% |
+
+| `src/core/export-html/template.html` | DIFFERENT | 54 | 76 | 43.1% | 56.9% |
+
+| `src/core/export-html/vendor/highlight.min.js` | DIFFERENT | 1212 | 1243 | 68.6% | 31.4% |
+
+| `src/core/export-html/vendor/marked.min.js` | DIFFERENT | 6 | 74 | 7.5% | 92.5% |
+
+| `src/cli.ts` | DIFFERENT | 15 | 39 | 40.7% | 59.3% |
+
+| `src/config.ts` | DIFFERENT | 241 | 202 | 80.4% | 19.6% |
+
+| `src/index.ts` | DIFFERENT | 332 | 294 | 6.7% | 93.3% |
+
+| `src/main.ts` | DIFFERENT | 870 | 826 | 57.0% | 43.0% |
+
+| `src/migrations.ts` | DIFFERENT | 295 | 206 | 12.4% | 87.6% |
+
+| `src/cli/args.ts` | DIFFERENT | 321 | 631 | 45.2% | 54.8% |
+
+| `src/cli/config-selector.ts` | DIFFERENT | 52 | 53 | 17.1% | 82.9% |
+
+| `src/cli/file-processor.ts` | DIFFERENT | 96 | 107 | 8.9% | 91.1% |
+
+| `src/cli/list-models.ts` | DIFFERENT | 104 | 101 | 13.7% | 86.3% |
+
+| `src/cli/session-picker.ts` | DIFFERENT | 51 | 93 | 8.3% | 91.7% |
+
+| `src/core/agent-session.ts` | DIFFERENT | 3182 | 3348 | 75.2% | 24.8% |
+
+| `src/core/auth-storage.ts` | DIFFERENT | 489 | 517 | 42.3% | 57.7% |
+
+| `src/core/bash-executor.ts` | DIFFERENT | 158 | 328 | 6.6% | 93.4% |
+
+| `src/core/diagnostics.ts` | DIFFERENT | 15 | 215 | 11.3% | 88.7% |
+
+| `src/core/event-bus.ts` | DIFFERENT | 33 | 37 | 28.6% | 71.4% |
+
+| `src/core/exec.ts` | DIFFERENT | 104 | 88 | 19.8% | 80.2% |
+
+| `src/core/footer-data-provider.ts` | DIFFERENT | 197 | 139 | 4.2% | 95.8% |
+
+| `src/core/index.ts` | DIFFERENT | 60 | 54 | 3.5% | 96.5% |
+
+| `src/core/keybindings.ts` | DIFFERENT | 220 | 159 | 13.2% | 86.8% |
+
+| `src/core/messages.ts` | DIFFERENT | 195 | 1 | 0.0% | 100.0% |
+
+| `src/core/model-registry.ts` | DIFFERENT | 718 | 695 | 64.1% | 35.9% |
+
+| `src/core/model-resolver.ts` | DIFFERENT | 628 | 315 | 7.4% | 92.6% |
+
+| `src/core/package-manager.ts` | DIFFERENT | 2047 | 808 | 3.4% | 96.6% |
+
+| `src/core/prompt-templates.ts` | DIFFERENT | 299 | 299 | 99.3% | 0.7% |
+
+| `src/core/resource-loader.ts` | DIFFERENT | 868 | 842 | 84.7% | 15.3% |
+
+| `src/core/sdk.ts` | DIFFERENT | 373 | 575 | 70.9% | 29.1% |
+
+| `src/core/session-manager.ts` | DIFFERENT | 1410 | 1521 | 93.2% | 6.8% |
+
+| `src/core/settings-manager.ts` | DIFFERENT | 953 | 745 | 61.7% | 38.3% |
+
+| `src/core/skills.ts` | DIFFERENT | 483 | 424 | 82.9% | 17.1% |
+
+| `src/core/system-prompt.ts` | DIFFERENT | 210 | 263 | 53.3% | 46.7% |
+
+| `src/core/timings.ts` | DIFFERENT | 25 | 25 | 32.0% | 68.0% |
+
+| `src/core/compaction/branch-summarization.ts` | DIFFERENT | 353 | 353 | 98.6% | 1.4% |
+
+| `src/core/compaction/compaction.ts` | DIFFERENT | 816 | 814 | 98.5% | 1.5% |
+
+| `src/core/compaction/utils.ts` | DIFFERENT | 170 | 155 | 92.9% | 7.1% |
+
+| `src/core/extensions/index.ts` | DIFFERENT | 166 | 141 | 87.9% | 12.1% |
+
+| `src/core/extensions/loader.ts` | DIFFERENT | 545 | 527 | 89.2% | 10.8% |
+
+| `src/core/extensions/runner.ts` | DIFFERENT | 908 | 650 | 79.5% | 20.5% |
+
+| `src/core/extensions/types.ts` | DIFFERENT | 1411 | 1142 | 87.0% | 13.0% |
+
+| `src/core/extensions/wrapper.ts` | DIFFERENT | 34 | 174 | 26.0% | 74.0% |
+
+| `src/core/tools/bash.ts` | DIFFERENT | 327 | 7 | 0.0% | 100.0% |
+
+| `src/core/tools/edit-diff.ts` | DIFFERENT | 309 | 13 | 3.7% | 96.3% |
+
+| `src/core/tools/edit.ts` | DIFFERENT | 227 | 7 | 0.0% | 100.0% |
+
+| `src/core/tools/find.ts` | DIFFERENT | 277 | 6 | 0.0% | 100.0% |
+
+| `src/core/tools/grep.ts` | DIFFERENT | 346 | 7 | 0.0% | 100.0% |
+
+| `src/core/tools/index.ts` | DIFFERENT | 140 | 473 | 26.1% | 73.9% |
+
+| `src/core/tools/ls.ts` | DIFFERENT | 170 | 7 | 0.0% | 100.0% |
+
+| `src/core/tools/path-utils.ts` | DIFFERENT | 94 | 5 | 0.0% | 100.0% |
+
+| `src/core/tools/read.ts` | DIFFERENT | 222 | 7 | 0.0% | 100.0% |
+
+| `src/core/tools/truncate.ts` | DIFFERENT | 265 | 10 | 0.0% | 100.0% |
+
+| `src/core/tools/write.ts` | DIFFERENT | 118 | 6 | 0.0% | 100.0% |
+
+| `src/modes/interactive/interactive-mode.ts` | DIFFERENT | 4556 | 6513 | 69.4% | 30.6% |
+
+| `src/modes/print-mode.ts` | DIFFERENT | 124 | 657 | 22.0% | 78.0% |
+
+| `src/modes/rpc/rpc-client.ts` | DIFFERENT | 505 | 1409 | 48.9% | 51.1% |
+
+| `src/modes/rpc/rpc-mode.ts` | DIFFERENT | 638 | 1367 | 51.0% | 49.0% |
+
+| `src/modes/rpc/rpc-types.ts` | DIFFERENT | 263 | 584 | 52.4% | 47.6% |
+
+| `src/modes/interactive/theme/theme.ts` | DIFFERENT | 1105 | 1524 | 83.5% | 16.5% |
+
+| `src/modes/interactive/theme/dark.json` | DIFFERENT | 85 | 85 | 98.8% | 1.2% |
+
+| `src/modes/interactive/theme/light.json` | DIFFERENT | 84 | 84 | 98.8% | 1.2% |
+
+| `src/modes/interactive/theme/theme-schema.json` | DIFFERENT | 335 | 335 | 99.4% | 0.6% |
+
+| `src/modes/interactive/components/armin.ts` | DIFFERENT | 382 | 383 | 99.6% | 0.4% |
+
+| `src/modes/interactive/components/assistant-message.ts` | DIFFERENT | 115 | 274 | 53.0% | 47.0% |
+
+| `src/modes/interactive/components/bash-execution.ts` | DIFFERENT | 210 | 507 | 57.7% | 42.3% |
+
+| `src/modes/interactive/components/bordered-loader.ts` | DIFFERENT | 66 | 67 | 97.7% | 2.3% |
+
+| `src/modes/interactive/components/branch-summary-message.ts` | DIFFERENT | 58 | 59 | 97.4% | 2.6% |
+
+| `src/modes/interactive/components/compaction-summary-message.ts` | DIFFERENT | 59 | 60 | 97.5% | 2.5% |
+
+| `src/modes/interactive/components/config-selector.ts` | DIFFERENT | 592 | 591 | 97.2% | 2.8% |
+
+| `src/modes/interactive/components/countdown-timer.ts` | DIFFERENT | 38 | 39 | 96.1% | 3.9% |
+
+| `src/modes/interactive/components/custom-editor.ts` | DIFFERENT | 80 | 82 | 93.8% | 6.2% |
+
+| `src/modes/interactive/components/custom-message.ts` | DIFFERENT | 99 | 108 | 92.8% | 7.2% |
+
+| `src/modes/interactive/components/dynamic-border.ts` | DIFFERENT | 25 | 26 | 94.1% | 5.9% |
+
+| `src/modes/interactive/components/extension-editor.ts` | DIFFERENT | 147 | 137 | 93.7% | 6.3% |
+
+| `src/modes/interactive/components/extension-input.ts` | DIFFERENT | 85 | 86 | 98.2% | 1.8% |
+
+| `src/modes/interactive/components/extension-selector.ts` | DIFFERENT | 107 | 108 | 98.6% | 1.4% |
+
+| `src/modes/interactive/components/footer.ts` | DIFFERENT | 216 | 235 | 78.5% | 21.5% |
+
+| `src/modes/interactive/components/index.ts` | DIFFERENT | 32 | 31 | 98.4% | 1.6% |
+
+| `src/modes/interactive/components/keybinding-hints.ts` | DIFFERENT | 66 | 67 | 97.7% | 2.3% |
+
+| `src/modes/interactive/components/login-dialog.ts` | DIFFERENT | 174 | 177 | 97.4% | 2.6% |
+
+| `src/modes/interactive/components/model-selector.ts` | DIFFERENT | 333 | 326 | 97.4% | 2.6% |
+
+| `src/modes/interactive/components/oauth-selector.ts` | DIFFERENT | 121 | 472 | 29.3% | 70.7% |
+
+| `src/modes/interactive/components/scoped-models-selector.ts` | DIFFERENT | 346 | 338 | 96.5% | 3.5% |
+
+| `src/modes/interactive/components/session-selector-search.ts` | DIFFERENT | 194 | 177 | 91.6% | 8.4% |
+
+| `src/modes/interactive/components/session-selector.ts` | DIFFERENT | 1019 | 835 | 82.5% | 17.5% |
+
+| `src/modes/interactive/components/settings-selector.ts` | DIFFERENT | 432 | 365 | 89.8% | 10.2% |
+
+| `src/modes/interactive/components/show-images-selector.ts` | DIFFERENT | 50 | 46 | 89.6% | 10.4% |
+
+| `src/modes/interactive/components/skill-invocation-message.ts` | DIFFERENT | 55 | 56 | 97.3% | 2.7% |
+
+| `src/modes/interactive/components/theme-selector.ts` | DIFFERENT | 67 | 63 | 92.3% | 7.7% |
+
+| `src/modes/interactive/components/thinking-selector.ts` | DIFFERENT | 74 | 65 | 87.8% | 12.2% |
+
+| `src/modes/interactive/components/tool-execution.ts` | DIFFERENT | 916 | 1133 | 66.3% | 33.7% |
+
+| `src/modes/interactive/components/tree-selector.ts` | DIFFERENT | 1194 | 1050 | 91.0% | 9.0% |
+
+| `src/modes/interactive/components/user-message-selector.ts` | DIFFERENT | 143 | 144 | 99.0% | 1.0% |
+
+| `src/modes/interactive/components/user-message.ts` | DIFFERENT | 33 | 204 | 13.5% | 86.5% |
+
+| `src/modes/interactive/components/visual-truncate.ts` | DIFFERENT | 50 | 51 | 97.0% | 3.0% |
+
+| `src/utils/clipboard-image.ts` | DIFFERENT | 300 | 157 | 63.0% | 37.0% |
+
+| `src/utils/clipboard.ts` | DIFFERENT | 81 | 53 | 46.3% | 53.7% |
+
+| `src/utils/frontmatter.ts` | DIFFERENT | 39 | 690 | 0.0% | 100.0% |
+
+| `src/utils/git.ts` | DIFFERENT | 192 | 6 | 2.0% | 98.0% |
+
+| `src/utils/image-convert.ts` | DIFFERENT | 41 | 38 | 93.7% | 6.3% |
+
+| `src/utils/image-resize.ts` | DIFFERENT | 235 | 232 | 98.1% | 1.9% |
+
+| `src/utils/shell.ts` | DIFFERENT | 202 | 180 | 91.1% | 8.9% |
+
+| `src/utils/tools-manager.ts` | DIFFERENT | 287 | 213 | 80.4% | 19.6% |
+
+| `src/core/discover-packages.ts` | MISSING_IN_ONE | 0 | 70 | 0.0% | 100.0% |
+
+| `src/core/subagents.ts` | MISSING_IN_ONE | 0 | 224 | 0.0% | 100.0% |
+
+| `src/core/task-session-manager.ts` | MISSING_IN_ONE | 0 | 521 | 0.0% | 100.0% |
+
+| `src/core/todo-store.ts` | MISSING_IN_ONE | 0 | 88 | 0.0% | 100.0% |
+
+| `src/core/tools/bg-process.ts` | MISSING_IN_ONE | 0 | 107 | 0.0% | 100.0% |
+
+| `src/core/tools/memory.ts` | MISSING_IN_ONE | 0 | 218 | 0.0% | 100.0% |
+
+| `src/core/tools/registry.ts` | MISSING_IN_ONE | 0 | 146 | 0.0% | 100.0% |
+
+| `src/core/tools/task.ts` | MISSING_IN_ONE | 0 | 107 | 0.0% | 100.0% |
+
+| `src/core/tools/todo.ts` | MISSING_IN_ONE | 0 | 136 | 0.0% | 100.0% |
+
+| `src/core/tools/webfetch.ts` | MISSING_IN_ONE | 0 | 6 | 0.0% | 100.0% |
+
+| `src/core/tools/websearch.ts` | MISSING_IN_ONE | 0 | 6 | 0.0% | 100.0% |
+
+| `src/modes/shared.ts` | MISSING_IN_ONE | 0 | 597 | 0.0% | 100.0% |
+
+| `src/observability/**` | MISSING_IN_ONE | 0 | 0 | 0.0% | 100.0% |
+
+| `src/utils/__tests__/**` | MISSING_IN_ONE | 0 | 0 | 0.0% | 100.0% |
+