@phren/cli 0.0.11 → 0.0.13

package/skills/docs.md

@@ -74,10 +74,10 @@ These numbers appear in multiple files and must all agree:
 
 | Item | Where to verify |
 |------|----------------|
-| MCP tool count | `mcp/src/index.ts` — count `server.tool(` calls |
-| CLI subcommand count | `mcp/src/cli.ts` — count `program.command(` calls |
+| MCP tool count | `mcp/src/index.ts`: count `server.tool(` calls |
+| CLI subcommand count | `mcp/src/cli.ts`: count `program.command(` calls |
 | Version | `package.json` → `version` field |
-| Hook events | `mcp/src/init.ts` — look for hook registration |
+| Hook events | `mcp/src/init.ts`: look for hook registration |
 
 Run `/parity` if it is installed to automate numeric cross-checking across surfaces.
 
@@ -95,7 +95,7 @@ Sections to verify and update:
 - **Tool count** in any summary sentence (e.g. "65 MCP tools")
 
 When updating the HTML:
-- Keep existing structure and CSS classes — do not restructure the page
+- Keep existing structure and CSS classes. Do not restructure the page
 - Update text content only; do not rewrite layout
 - Preserve the `<details>` blocks for env var categories if they exist
 
@@ -103,12 +103,12 @@ When updating the HTML:
 
 These files are consumed by LLMs directly. Keep them plain text with no HTML.
 
-`llms.txt` — short summary (under 60 lines):
+`llms.txt`, short summary (under 60 lines):
 - Tool count
 - Install command
 - One-line description of what phren does
 
-`llms-full.txt` — comprehensive reference:
+`llms-full.txt`, comprehensive reference:
 - All MCP tool signatures with descriptions
 - All CLI commands
 - All environment variables
@@ -122,7 +122,7 @@ The CLI commands block must list every top-level command. Add or remove lines to
 
 ### 7. Sync version references
 
-`CHANGELOG.md`: the top entry's version must match `package.json`. If a new version was bumped but the changelog has no entry yet, note it — do not fabricate one.
+`CHANGELOG.md`: the top entry's version must match `package.json`. If a new version was bumped but the changelog has no entry yet, note it. Do not fabricate one.
 
 `README.md` and `docs/index.html`: update any hardcoded version strings (badges, `npm install @phren/cli@X.Y.Z`, etc.).
 
@@ -150,18 +150,18 @@ No changes needed:
   - mcp/README.md: tool count correct
 
 Warnings:
-  - docs/faq.md: references `phren init` — command was renamed to `phren link` in v0.0.7
+  - docs/faq.md: references `phren init`, command was renamed to `phren link` in v0.0.7
 ```
 
 List every file checked. Be specific about what changed and what line/section.
 
 ## What not to do
 
-- Do not restructure or reformat documentation for style — only fix accuracy
+- Do not restructure or reformat documentation for style. Only fix accuracy
 - Do not add new sections or features to the docs without user direction
 - Do not fabricate changelog entries for unreleased versions
-- Do not silently skip a surface because it looks "probably fine" — check all of them
-- Do not guess tool counts — always derive from `grep` on the source file
+- Do not silently skip a surface because it looks "probably fine". Check all of them
+- Do not guess tool counts. Always derive from `grep` on the source file
 
 ## Related skills
 

package/starter/README.md

@@ -1,6 +1,6 @@
 # My Phren
 
-Your personal project store for [phren](https://github.com/alaarab/phren). Phren is your project's memory keeper — he holds what your agents learn and surfaces it when it matters, across sessions and machines.
+Your personal project store for [phren](https://github.com/alaarab/phren). Phren is your project's memory keeper. He holds what your agents learn and surfaces it when it matters, across sessions and machines.
 
 ## Structure
 

package/starter/global/CLAUDE.md

@@ -56,7 +56,7 @@ Run `/phren-sync` to pull everything down or push changes back.
 
 ## MCP tools
 
-The phren MCP server is running. Phren already knows a lot — ask him before asking the user to repeat themselves.
+The phren MCP server is running. Phren already knows a lot. Ask him before asking the user to repeat themselves.
 
 - **At session start:** ask phren what's active: `list_projects()`, then `get_project_summary(name)` for the relevant project
 - **When the user mentions a project, codebase, or task:** ask phren first: `search_knowledge(query)` before asking questions
@@ -86,4 +86,4 @@ Read `~/.phren-context.md` at the start of every session for machine-specific co
 
 ## Without MCP server
 
-If the MCP server isn't available, phren still helps. Claude reads `~/.phren-context.md` and per-project memory files from `~/.claude/projects/` for context, then fetches details directly from `~/.phren/project-name/` as needed. No MCP required — phren just works a bit more quietly.
+If the MCP server isn't available, phren still helps. Claude reads `~/.phren-context.md` and per-project memory files from `~/.claude/projects/` for context, then fetches details directly from `~/.phren/project-name/` as needed. No MCP required, phren just works a bit more quietly.

package/starter/global/skills/audit.md

@@ -1,6 +1,6 @@
 ---
 name: audit
-description: Full codebase audit — dead code, security, dependencies, performance, optimization. Not a diff review — scans everything.
+description: Full codebase audit. Dead code, security, dependencies, performance, optimization. Not a diff review, scans everything.
 ---
 # /audit - Full Codebase Audit
 
@@ -15,7 +15,7 @@ Launch 5 parallel agents. Each one scans the full codebase for a different class
 Before launching agents, understand the project:
 
 ```
-1. Read package.json (or pyproject.toml, Cargo.toml, go.mod — whatever applies)
+1. Read package.json (or pyproject.toml, Cargo.toml, go.mod, whatever applies)
 2. Find the source directories (src/, lib/, app/, etc.)
 3. Count files by extension to understand the stack
 4. Check for existing lint/test configs
@@ -32,7 +32,7 @@ Use the Agent tool to launch all five concurrently in a single message. Give eac
 Find code that exists but isn't used:
 
 1. **Unused exports.** For every `export` in the codebase, check if it's imported anywhere. Flag exports that are only used in their own file or not used at all. Exclude entry points and public API surfaces.
-2. **Unused dependencies.** Cross-reference `package.json` dependencies against actual imports in source files. Flag packages that are installed but never imported. Check devDependencies too — are test utilities actually used in tests?
+2. **Unused dependencies.** Cross-reference `package.json` dependencies against actual imports in source files. Flag packages that are installed but never imported. Check devDependencies too. Are test utilities actually used in tests?
 3. **Dead functions.** Functions defined but never called. Methods on classes that nothing invokes. Event handlers registered but for events that are never emitted.
 4. **Orphan files.** Files that nothing imports. Test files for source files that no longer exist. Config files for tools that aren't in the project.
 5. **Feature flags that resolved.** Environment variable checks where one branch is clearly dead. TODO/FIXME/HACK comments older than 6 months.
@@ -91,16 +91,16 @@ Wait for all agents. Then:
 
 1. **Deduplicate.** Multiple agents may flag the same issue from different angles. Merge them.
 2. **Prioritize.** Security issues first. Then dead code (easy wins). Then performance. Then quality.
-3. **Fix directly.** Don't just report — fix what you can. For things that need the user's input (like removing a dependency that might be used in a way you can't see), ask.
+3. **Fix directly.** Don't just report. Fix what you can. For things that need the user's input (like removing a dependency that might be used in a way you can't see), ask.
 4. **Summarize.** Report what was found, what was fixed, and what needs the user's decision.
 
 ## Options
 
 The user can scope the audit:
 
-- `/audit` — full audit, all 5 agents
-- `/audit security` — just the security agent
-- `/audit dead-code` — just dead code detection
-- `/audit performance` — just performance
-- `/audit deps` — just dependency health
-- `/audit quality` — just code quality
+- `/audit`: full audit, all 5 agents
+- `/audit security`: just the security agent
+- `/audit dead-code`: just dead code detection
+- `/audit performance`: just performance
+- `/audit deps`: just dependency health
+- `/audit quality`: just code quality

package/mcp/dist/cli-hooks-retrieval.js

@@ -1,2 +0,0 @@
-// cli-hooks-retrieval.ts — re-export surface for the shared retrieval core.
-export * from "./shared-retrieval.js";

package/mcp/dist/impact-scoring.js

@@ -1,22 +0,0 @@
-// Backward-compatible wrapper around finding-impact.
-import { findingIdFromLine, extractFindingIdsFromSnippet, logImpact, getHighImpactFindings, markImpactEntriesCompletedForSession, } from "./finding-impact.js";
-export function impactEntryKey(project, findingId) {
-    return `${project}\u0000${findingId}`;
-}
-export { findingIdFromLine, extractFindingIdsFromSnippet, markImpactEntriesCompletedForSession, };
-export function appendImpactEntries(phrenPath, entries) {
-    const pending = entries.filter((entry) => !entry.taskCompleted);
-    if (pending.length === 0)
-        return;
-    logImpact(phrenPath, pending.map((entry) => ({
-        findingId: entry.findingId,
-        project: entry.project,
-        sessionId: entry.sessionId,
-    })));
-}
-export function getHighImpactFindingKeys(phrenPath, minSuccessCount = 3) {
-    const findingIds = getHighImpactFindings(phrenPath, minSuccessCount);
-    // Legacy API encoded project+findingId; new API tracks finding ID globally.
-    // Return IDs as-is to preserve compatibility where only membership checks are used.
-    return findingIds;
-}