prizmkit 1.1.79 → 1.1.80

package/bundled/skills-windows/app-planner/references/project-conventions-discovery.md

@@ -0,0 +1,59 @@
+# Project Conventions Discovery
+
+AI-driven procedure for discovering and capturing project-level conventions. SKILL.md retains the trigger condition and entry point; this file contains the full Analyze → Reason → Present workflow.
+
+## Principle
+
+**Do NOT follow any fixed checklist.** Every project is different. You must analyze the project first, then reason about what system-level conventions this specific project needs.
+
+## Step 1: Analyze the project
+
+- Read tech stack, dependencies, existing code, config files, README, any existing style guides or linter configs
+- For brownfield: also read actual source code patterns (naming, file structure, error handling, etc.)
+- For greenfield: use the user's stated goals and intended tech stack
+
+## Step 2: Reason about what conventions matter for THIS project
+
+Think about what decisions, if left unstandardized, would cause inconsistency as the project grows. Consider all dimensions relevant to the project — these might include (but are NOT limited to):
+
+- Language and localization choices
+- Code style and naming patterns
+- Architecture and communication patterns
+- Data format and storage decisions
+- Security and auth approaches
+- UI/UX patterns
+- Testing strategies
+- Deployment and environment patterns
+- ...anything else you observe that needs a project-level decision
+
+The point is: YOU decide what's relevant based on what you see. A Next.js SaaS app needs completely different conventions than a Python data pipeline or a Go microservice.
+
+## Step 3: Present findings via `AskUserQuestion`
+
+First, show "Already decided" conventions as text:
+> **Already decided** (detected from your codebase):
+> - [convention]: [value] (source: [where you found it])
+> - [convention]: [value] (source: [where you found it])
+
+Then use `AskUserQuestion` for conventions that need user input (up to 4 questions per call, use multiple calls as needed — no limit on total rounds). Each question:
+- Question text includes the convention name AND why it matters for this project
+- Options are the reasonable choices (2-4 per question)
+- Mark the recommended option first with "(Recommended)" in its label
+- Use `description` field to explain trade-offs
+
+After each batch of `AskUserQuestion` calls, reassess: are there more project-level conventions to cover? If yes, continue with more `AskUserQuestion` calls. Keep going until ALL project-level conventions are fully addressed.
+
+Then ask in text: "Anything I missed that you'd like to standardize?" — if the user adds more, continue the discovery loop.
+
+## Rules
+
+- **No interaction limit** — keep asking until every project-level convention is covered. Do NOT stop early or batch-skip to save rounds.
+- Every proposed convention must be justified by something you observed in the project — explain WHY it matters
+- Auto-confirm anything already evident from the codebase (show as "detected", let user override)
+- Propose as many conventions as the project genuinely needs, but don't pad with irrelevant ones
+- The "Anything I missed?" question is NOT the end — if the user adds items, ask follow-up `AskUserQuestion` calls to clarify those too
+
+## After Discovery
+
+→ Save answers to `AGENTS.md` / `CLAUDE.md` / `CODEBUDDY.md` under `### Project Conventions` section (format: one bullet per convention)
+→ Output format will naturally vary per project — that is the intended behavior

package/bundled/skills-windows/app-planner/references/project-state-detection.md

@@ -0,0 +1,90 @@
+# Project State Detection — Brownfield Behavior
+
+Detailed procedures for detecting and handling existing (brownfield) projects during app planning.
+
+## Detection Signals
+
+| Signal | Greenfield | Brownfield |
+|--------|-----------|------------|
+| `package.json` / `pyproject.toml` / `go.mod` / `Cargo.toml` / `pom.xml` | absent | present |
+| `src/` or `app/` directory with source files | absent | present |
+| `.git` with commit history | absent or initial commit only | present with history |
+| Empty or near-empty directory | yes | no |
+
+## Greenfield Behavior (default)
+
+Proceed with the standard Core Workflow — ask all questions from scratch.
+
+## Brownfield Behavior
+
+When an existing project is detected:
+
+### Step 1: Prerequisite Check (Mandatory)
+
+Before ANY planning work, check if AI-essential project context files exist:
+
+| File | Purpose | Status |
+|------|---------|--------|
+| `.prizmkit/prizm-docs/root.prizm` | Project architecture context for AI | exists / missing |
+| `.prizmkit/config.json` | Tech stack + runtime config | exists / missing |
+| `.prizmkit/plans/project-brief.md` | Product vision checklist | exists / missing |
+
+**If ANY are missing**, show the status table, then use `AskUserQuestion`:
+
+**Question**: "Some AI context files are missing. These help AI understand your project — making planning much more effective. How would you like to proceed?"
+- **Run project init first (Recommended)** — invoke `prizmkit-init` to scan your codebase and generate these files, then return to planning
+- **Continue without init** — I'll scan the project manually during this session (less thorough)
+- **Skip, I'll set these up later** — proceed with planning using only what's available
+
+- **Run project init first** -> Invoke `prizmkit-init`, then resume app-planner from where it left off
+- **Continue without init** -> Continue with Step 2 below (manual scan)
+- **Skip** -> Continue with Step 3, skip scanning
+
+### Step 2: Proactive Project Scanning
+
+Do NOT ask the user to describe their project — read it yourself first:
+
+1. **Scan project structure** to understand the codebase layout:
+   ```powershell
+   Get-ChildItem -Path . -Directory -Recurse -Depth 2 |
+     Where-Object { $_.FullName -notmatch '\\(node_modules|\.git|dist|build|__pycache__|vendor)(\\|$)' } |
+     ForEach-Object { $_.FullName -replace '[^\\]*\\', '|____' -replace '____\|', ' |' }
+   ```
+
+2. **Read existing project metadata** to infer tech stack and purpose:
+   - `package.json` -> name, description, dependencies, scripts
+   - `pyproject.toml` / `requirements.txt` -> Python dependencies
+   - `go.mod` -> Go module info
+   - `README.md` -> project description and goals
+   - `.prizmkit/config.json` -> previously detected tech stack
+   - `.prizmkit/prizm-docs/root.prizm` -> existing architecture context
+
+3. **Read key source files** (entry points, main routes, core models) to understand what the project actually does — don't rely solely on metadata.
+
+### Step 3: Present Inferred Summary with Confirmation
+
+Show the summary as text, then use `AskUserQuestion`:
+
+> Based on my analysis of your codebase:
+>
+> **Project**: [name] — [inferred description]
+> **Tech Stack**: [framework] + [language] + [key dependencies]
+> **Key Features Found**: [list 3-5 detected capabilities]
+> **Architecture**: [e.g., monolithic, microservices, serverless]
+
+**Question**: "Does this look correct?"
+- **Yes, looks correct (Recommended)** — proceed with planning
+- **Mostly correct, with changes** — I'll note corrections
+- **This is off** — let me describe the project
+
+### Step 4: Pre-fill and Focus
+
+- Phase 2 tech stack selection -> largely pre-filled from dependencies
+- Vision/problem statement -> inferred from README or package description (user confirms)
+- Existing features -> note them as `[x]` items in project brief
+
+**Focus remaining questions** (as options where possible) on what CANNOT be inferred:
+- Target users and core value proposition
+- Future direction and planned capabilities
+- Non-functional requirements (performance, scale, security)
+- Design direction (for frontend projects)

package/bundled/skills-windows/app-planner/references/rules-configuration.md

@@ -0,0 +1,46 @@
+# Rules Configuration — Detailed Q&A Procedures
+
+Procedural details for configuring per-layer development rules during app planning. SKILL.md retains Steps 1-3 and 5 (flow control); this file contains Step 4's full Q&A workflow.
+
+## Step 4: Configure Each Selected Layer — Full Q&A Workflow
+
+For each selected layer, run the 4-phase rule generation pattern:
+
+### Phase A — Load Layer Resources
+
+- Read `${SKILL_DIR}/references/rules/<layer>/fixed-rules.md` — industry-consensus rules injected without asking
+- Read `${SKILL_DIR}/references/rules/<layer>/question-bank.md` — interactive questions organized in groups (G1->G10): authoritative source for question text, options, "Recommended" markers, and Notes
+- Read `${SKILL_DIR}/references/rules/<layer>/question-manifest.json` — machine-readable structure for this layer. **This is your asking checklist.** It lists, per question: `group`, `required`, `maps_to` (which template placeholders the answer fills), and conditional fields (`required_if`, `auto_derived_when`, `options_vary_by`). Its `groups[]` carries each group's `quick_mode` flag, and `template_placeholders` is the expected-set for the Phase D self-check. Use the manifest to track coverage; use question-bank.md for the actual wording you present to the user. If a question id in one file is absent from the other, trust question-bank.md for content and note the drift.
+
+### Phase B — Interactive Q&A
+
+- Ask questions one group at a time (max 3 questions per message), as defined in question-bank.md
+- Each question shows a "Recommended" option to reduce decision cost
+- **Quick mode**: ask only the groups whose `quick_mode` is `true` in `question-manifest.json`. All other groups adopt recommended defaults silently. (The manifest is the single source of truth for quick-mode membership — do not hardcode group lists here.)
+- **Full mode**: ask all groups in manifest order (`groups[]`).
+- **Conditional questions** (driven by the manifest's per-question fields):
+  - `required_if: "<expr>"` — ask this question only when the expression over prior answers holds (e.g. Q14 `required_if: Q13 != D`). If the condition is false, the question is legitimately skipped — not a coverage gap.
+  - `auto_derived_when: "<expr>"` — when this holds, fill the mapped placeholder(s) from the prior answer without asking (e.g. backend Q6 `auto_derived_when: Q5 in [GraphQL, gRPC]`). Count as satisfied, not skipped.
+  - `options_vary_by: "<Qid>"` — the question is still asked; only its option list depends on that prior answer (e.g. framework/ORM options vary by language). Never affects whether the question is required.
+- Shortcut commands work at any point:
+  - `recommended` / `default` -> adopt all recommended for current group
+  - `all recommended` -> adopt all recommended for all remaining groups
+  - `skip` -> mark current group as "Not required at this stage"
+- Record answers in memory after each group. Track which manifest questions are answered, auto-derived, or conditionally skipped so Phase D can verify coverage.
+
+### Phase C — Auto-derivation
+
+- Read `${SKILL_DIR}/references/rules/<layer>/derivation-rules.md`
+- Match user answers against the trigger map using keyword matching
+- Derive platform-specific rules without asking the user (e.g., choosing Flutter -> auto-inject Flutter widget/persistence/navigation rules)
+
+### Phase D — Render and Write
+
+- Read `${SKILL_DIR}/references/rules/<layer>/template.md`
+- Fill all template placeholders with accumulated content from Phases A+B+C
+- **Post-render self-check** (driven by `question-manifest.json` -> `template_placeholders`):
+  1. **Coverage pass** — for every placeholder in `from_questions`, confirm it traces to an answered question OR an `auto_derived_when` path OR a conditionally-skipped question (`required_if` false). If a `from_questions` placeholder has no source, you skipped a required question — go back and ask it before writing.
+  2. **Residual pass** — scan the rendered document for any residual `{{ ` or ` }}`; count must be 0. Placeholders in `from_fixed_rules` / `auto_generated` / `metadata` are filled from fixed-rules.md, Phase D generation, and project metadata respectively — they are NOT expected to come from Q&A, but they MUST still be rendered (no residual braces).
+- Generate Appendix A (Deny List) and Appendix B (Recommended Tools) per template instructions — these fill the `auto_generated` placeholders
+- Create `.prizmkit/rules/` directory if it doesn't exist
+- Write `.prizmkit/rules/<layer>-rules.md`

package/bundled/skills-windows/bug-fix-workflow/SKILL.md

@@ -95,36 +95,7 @@ For trivial bugs with clear root cause and minimal scope:
 
 #### Step 1.2: Systematic Bug Clarification
 
-Ask questions across these dimensions until every aspect is clear. **Adapt to what the user has already provided** — skip questions that are already answered.
-
-**Reproduction Conditions:**
-- What exact steps trigger the bug? (step-by-step)
-- Which environment/browser/OS/version?
-- Is it reproducible every time, or intermittent?
-- When did it first appear? (after a specific change/deploy?)
-- Does it happen for all users or only specific accounts/roles/data?
-
-**Expected vs Actual Behavior:**
-- What should happen? (the correct behavior)
-- What actually happens? (the buggy behavior)
-- Is there partial functionality (e.g., works for some inputs but not others)?
-
-**Scope and Impact:**
-- Which features/pages/modules are affected?
-- Are there workarounds users are currently using?
-- Is this blocking other work?
-- Are there related symptoms elsewhere?
-
-**Data and State:**
-- What data/state triggers the issue? (specific input values, DB state, user session state)
-- Does the bug involve data corruption or just incorrect display/behavior?
-- If database-related: which tables/records are affected?
-
-**Error Details** (if not already provided):
-- Full error message and stack trace?
-- Browser console errors?
-- Server-side logs?
-- Network request/response details?
+Read `${SKILL_DIR}/references/bug-diagnosis.md` for the full question framework — reproduction conditions, expected vs actual behavior, scope/impact, data/state, and error details. Adapt to what the user has already provided; skip answered questions.
 
 #### Step 1.3: Confirmation Before Triage
 

package/bundled/skills-windows/bug-fix-workflow/references/bug-diagnosis.md

@@ -0,0 +1,66 @@
+# Bug Diagnosis — Question Framework
+
+Procedural details for systematic bug clarification in bug-fix-workflow Phase 1.
+
+## Step 1.2: Systematic Bug Clarification
+
+Ask questions across these dimensions until every aspect is clear. **Adapt to what the user has already provided** — skip questions that are already answered.
+
+### Reproduction Conditions
+- What exact steps trigger the bug? (step-by-step)
+- Which environment/browser/OS/version?
+- Is it reproducible every time, or intermittent?
+- When did it first appear? (after a specific change/deploy?)
+- Does it happen for all users or only specific accounts/roles/data?
+
+### Expected vs Actual Behavior
+- What should happen? (the correct behavior)
+- What actually happens? (the buggy behavior)
+- Is there partial functionality (e.g., works for some inputs but not others)?
+
+### Scope and Impact
+- Which features/pages/modules are affected?
+- Are there workarounds users are currently using?
+- Is this blocking other work?
+- Are there related symptoms elsewhere?
+
+### Data and State
+- What data/state triggers the issue? (specific input values, DB state, user session state)
+- Does the bug involve data corruption or just incorrect display/behavior?
+- If database-related: which tables/records are affected?
+
+### Error Details (if not already provided)
+- Full error message and stack trace?
+- Browser console errors?
+- Server-side logs?
+- Network request/response details?
+
+## Complexity Assessment
+
+### Simple bug → Fast Path candidate (ALL must be true)
+- Root cause is immediately obvious (typo, missing null check, wrong variable name, off-by-one)
+- Fix is ≤10 lines of code in a single file
+- No cross-module impact
+- Existing tests cover the affected path (or bug is in untested utility)
+- No data model or API changes
+
+### Complex bug → Planning Path (ANY is true)
+- Cross-module impact (>2 files affected)
+- Data model or API changes required
+- Root cause is uncertain or multi-layered
+- Fix requires structural changes
+- Multiple interrelated symptoms
+
+## Comparison with Pipeline Bug Fix
+
+| Dimension | bug-fix-workflow | bugfix-pipeline-launcher |
+|-----------|-------------------|----------------------------|
+| Scope | One bug at a time | All bugs in batch |
+| Execution | Interactive, in-session | Foreground or background daemon |
+| Diagnosis | Deep interactive Q&A with user | Automated from bug description |
+| Branch | Creates `fix/<BUG_ID>-*` branch | Pipeline manages branches |
+| Visibility | Full user interaction at each phase | Async, check status periodically |
+| User verification | Yes (Phase 6) | No (automated) |
+| Best for | Complex bugs needing user input | Batch of well-defined bugs |
+| Artifacts | Same (fix-plan.md + fix-report.md) | Same |
+| Commit prefix | `fix(<scope>):` | `fix(<scope>):` |

package/bundled/skills-windows/bugfix-pipeline-launcher/SKILL.md

@@ -201,23 +201,7 @@ Detect user intent from their message, then follow the corresponding workflow:
    - xhigh — Extensive reasoning (`PRIZMKIT_EFFORT=xhigh`)
    - max — Maximum reasoning, Claude Code only (`PRIZMKIT_EFFORT=max`)
 
-   **Environment variable mapping** (for translating user responses → env vars):
-
-   | Config choice | Environment variable |
-   |-----------|---------------------|
-   | Verbose: Off | `VERBOSE=0` |
-   | Verbose: On | `VERBOSE=1` |
-   | Max retries: N | `MAX_RETRIES=N` |
-   | Critic: On | `ENABLE_CRITIC=true` |
-   | Timeout: value | `SESSION_TIMEOUT=<seconds>` |
-   | Stop on failure: On | `STOP_ON_FAILURE=1` |
-   | Effort: value | `PRIZMKIT_EFFORT=<value>` |
-
-   **Advanced environment variables** (not exposed in interactive menu, pass via `--env`):
-
-   | Variable | Default | Purpose |
-   |----------|---------|---------|
-   | `MODEL` | (none) | AI model override (e.g. `claude-opus-4.6`) |
+   **Environment variable mapping** and **advanced environment variables** tables — read `${SKILL_DIR}/references/configuration.md` for the full translation tables (8 config→env mappings + 8 advanced variables).
 
    ⚠️ STOP HERE and wait for user response before continuing to step 6.
 
@@ -364,18 +348,7 @@ When user says "retry B-001":
 
 ### Error Handling
 
-| Error | Action |
-|-------|--------|
-| `.prizmkit/plans/bug-fix-list.json` not found | Tell user to run `bug-planner` skill first |
-| JSON parsing failed | Use the bundled Python validation/status scripts instead of external JSON tools |
-| AI CLI not in PATH | Check Codex (`codex`), Claude (`claude`), or CodeBuddy (`cbc`) installation |
-| Bugfix pipeline already running | Show status, ask if user wants to stop and restart |
-| PID file stale (process dead) | `launch-bugfix-daemon.ps1` auto-cleans, retry start |
-| Launch failed (process died immediately) | Show last 20 lines of log: `Get-Content -Tail 20 .prizmkit/state/daemon/bugfix-daemon.log` |
-| All bugs blocked/failed/needs-info | Show status, suggest retrying or providing more info |
-| `playwright-cli` not installed | Browser verification skipped for playwright bugs (non-blocking). Suggest: `npm install -g @playwright/cli@latest; playwright-cli install --skills` |
-| `opencli` not installed | Browser verification skipped for opencli bugs (non-blocking). Install opencli for Chrome session-based browser verification |
-| PowerShell execution policy blocks script | Run `Set-ExecutionPolicy -Scope Process Bypass` for the current terminal |
+Read `${SKILL_DIR}/references/configuration.md` for the full error handling table (11 entries).
 
 ### Integration Notes
 

package/bundled/skills-windows/bugfix-pipeline-launcher/references/configuration.md

@@ -0,0 +1,49 @@
+# Configuration Reference — Bugfix Pipeline Launcher (Windows)
+
+Environment variable mappings for the bugfix launcher.
+
+## Environment Variable Mapping
+
+Translating user responses to env vars:
+
+| Config choice | Environment variable |
+|-----------|---------------------|
+| Verbose: Off | `VERBOSE=0` |
+| Verbose: On | `VERBOSE=1` |
+| Max retries: N | `MAX_RETRIES=N` |
+| Critic: On | `ENABLE_CRITIC=true` |
+| Timeout: value | `SESSION_TIMEOUT=<seconds>` |
+| Stop on failure: On | `STOP_ON_FAILURE=1` |
+| Deploy: Yes | `ENABLE_DEPLOY=1` |
+| Effort: value | `PRIZMKIT_EFFORT=<value>` |
+
+## Advanced Environment Variables
+
+Not exposed in interactive menu, pass via `--env`:
+
+| Variable | Default | Purpose |
+|----------|---------|---------|
+| `MODEL` | (none) | AI model override (e.g. `claude-opus-4.6`) |
+| `AUTO_PUSH` | `0` | Auto-push to remote after successful bug fix (`1` to enable) |
+| `DEV_BRANCH` | auto-generated | Custom dev branch name (default: `bugfix/pipeline-{run_id}`) |
+| `HEARTBEAT_INTERVAL` | `30` | Heartbeat log interval in seconds |
+| `HEARTBEAT_STALE_THRESHOLD` | `600` | Max seconds without heartbeat before marking stale |
+| `LOG_CLEANUP_ENABLED` | `1` | Run periodic log cleanup (`0` to disable) |
+| `LOG_RETENTION_DAYS` | `14` | Delete logs older than N days |
+| `LOG_MAX_TOTAL_MB` | `1024` | Keep total logs under N MB via oldest-first cleanup |
+
+## Error Handling
+
+| Error | Action |
+|-------|--------|
+| `.prizmkit/plans/bug-fix-list.json` not found | Tell user to run `bug-planner` skill first |
+| JSON parsing failed | Use the bundled Python validation/status scripts instead of external JSON tools |
+| AI CLI not in PATH | Check Codex (`codex`), Claude (`claude`), or CodeBuddy (`cbc`) installation |
+| Bugfix pipeline already running | Show status, ask if user wants to stop and restart |
+| PID file stale (process dead) | `launch-bugfix-daemon.ps1` auto-cleans, retry start |
+| Launch failed (process died immediately) | Show last 20 lines of log: `Get-Content -Tail 20 .prizmkit/state/daemon/bugfix-daemon.log` |
+| All bugs blocked/failed/needs-info | Show status, suggest retrying or providing more info |
+| `playwright-cli` not installed | Browser verification skipped for playwright bugs (non-blocking). Suggest: `npm install -g @playwright/cli@latest; playwright-cli install --skills` |
+| `opencli` not installed | Browser verification skipped for opencli bugs (non-blocking). Install opencli for Chrome session-based browser verification |
+| Deploy session failed | Pipeline completed but deploy session exited non-zero. Check `.prizmkit/state/bugfix/deploy/<session_id>/logs/session.log`. Retry manually: `/prizmkit-deploy`. |
+| PowerShell execution policy blocks script | Run `Set-ExecutionPolicy -Scope Process Bypass` for the current terminal |

package/bundled/skills-windows/feature-pipeline-launcher/SKILL.md

@@ -219,24 +219,7 @@ Detect user intent from their message, then follow the corresponding workflow:
    - xhigh — Extensive reasoning (`PRIZMKIT_EFFORT=xhigh`)
    - max — Maximum reasoning, Claude Code only (`PRIZMKIT_EFFORT=max`)
 
-   **Environment variable mapping** (for translating user responses → env vars):
-
-   | Config choice | Environment variable |
-   |-----------|---------------------|
-   | Critic: On | `ENABLE_CRITIC=true` |
-   | Verbose: Off | `VERBOSE=0` |
-   | Verbose: On | `VERBOSE=1` |
-   | Max retries: N | `MAX_RETRIES=N` |
-   | Timeout: value | `SESSION_TIMEOUT=<seconds>` |
-   | Stop on failure: On | `STOP_ON_FAILURE=1` |
-   | Effort: value | `PRIZMKIT_EFFORT=<value>` |
-
-   **Advanced environment variables** (not exposed in interactive menu, pass via `--env`):
-
-   | Variable | Default | Purpose |
-   |----------|---------|---------|
-   | `MODEL` | (none) | AI model override (e.g. `claude-opus-4.6`) |
-   | `PIPELINE_MODE` | (none) | Override mode for all features: `lite`\|`standard`\|`full` |
+   **Environment variable mapping** and **advanced environment variables** tables — read `${SKILL_DIR}/references/configuration.md` for the full translation tables (8 config→env mappings + 9 advanced variables).
 
    ⚠️ STOP HERE and wait for user response before continuing to step 7.
 
@@ -422,23 +405,7 @@ After pipeline completion, if features have `browser_interaction` fields and the
 
 ### Error Handling
 
-| Error | Action |
-|-------|--------|
-| `.prizmkit/plans/feature-list.json` not found | Tell user to run `feature-planner` skill first |
-| JSON parsing failed | Use the bundled Python validation/status scripts instead of external JSON tools |
-| AI CLI not in PATH | Check Codex (`codex`), Claude (`claude`), or CodeBuddy (`cbc`) installation |
-| Pipeline already running | Show status, ask if user wants to stop and restart |
-| PID file stale (process dead) | `launch-feature-daemon.ps1` auto-cleans, retry start |
-| Launch failed (process died immediately) | Show last 20 lines of log: `Get-Content -Tail 20 .prizmkit/state/daemon/feature-daemon.log` |
-| Feature stuck/blocked | Use `reset-feature.ps1 <F-XXX> --clean --run` for a fresh retry |
-| All features blocked/failed | Show status, suggest daemon-safe recovery: `.\.prizmkit\dev-pipeline\reset-feature.ps1 <F-XXX> --clean --run .prizmkit/plans/feature-list.json` |
-| `playwright-cli` not installed | Browser verification skipped for playwright features (non-blocking). Suggest: `npm install -g @playwright/cli@latest; playwright-cli install --skills` |
-| `opencli` not installed | Browser verification skipped for opencli features (non-blocking). Install opencli for Chrome session-based browser verification |
-| PowerShell execution policy blocks script | Run `Set-ExecutionPolicy -Scope Process Bypass` for the current terminal |
-| Pipeline stop failed | Use `Stop-Process -Id <PID> -Force`; inspect AI CLI processes with `Get-Process | Select-String codex` |
-| `.env.local` missing or incomplete | Warn: database connection variables not found. Suggest creating env file with required connection variables for the project's database |
-| Database unreachable | Warn: database features will produce mock-only tests. Suggest checking database service status and connection credentials |
-| Migrations not applied | Warn: tables or schema referenced in migration files not found in database. Suggest applying pending migrations |
+Read `${SKILL_DIR}/references/configuration.md` for the full error handling table (16 entries covering feature-list not found, dependency checks, pipeline state, browser tooling, permission, and deploy failures).
 
 ### Integration Notes
 

package/bundled/skills-windows/feature-pipeline-launcher/references/configuration.md

@@ -0,0 +1,55 @@
+# Configuration Reference — Feature Pipeline Launcher (Windows)
+
+Environment variable mappings for the feature launcher.
+
+## Environment Variable Mapping
+
+Translating user responses to env vars:
+
+| Config choice | Environment variable |
+|-----------|---------------------|
+| Critic: On | `ENABLE_CRITIC=true` |
+| Verbose: Off | `VERBOSE=0` |
+| Verbose: On | `VERBOSE=1` |
+| Max retries: N | `MAX_RETRIES=N` |
+| Timeout: value | `SESSION_TIMEOUT=<seconds>` |
+| Stop on failure: On | `STOP_ON_FAILURE=1` |
+| Deploy: Yes | `ENABLE_DEPLOY=1` |
+| Effort: value | `PRIZMKIT_EFFORT=<value>` |
+
+## Advanced Environment Variables
+
+Not exposed in interactive menu, pass via `--env`:
+
+| Variable | Default | Purpose |
+|----------|---------|---------|
+| `MODEL` | (none) | AI model override (e.g. `claude-opus-4.6`) |
+| `AUTO_PUSH` | `0` | Auto-push to remote after successful feature (`1` to enable) |
+| `DEV_BRANCH` | auto-generated | Custom dev branch name (default: `dev/{feature_id}-YYYYMMDDHHmm`) |
+| `HEARTBEAT_INTERVAL` | `30` | Heartbeat log interval in seconds |
+| `HEARTBEAT_STALE_THRESHOLD` | `600` | Max seconds without heartbeat before marking stale |
+| `PIPELINE_MODE` | (none) | Override mode for all features: `lite`\|`standard`\|`full` |
+| `LOG_CLEANUP_ENABLED` | `1` | Run periodic log cleanup (`0` to disable) |
+| `LOG_RETENTION_DAYS` | `14` | Delete logs older than N days |
+| `LOG_MAX_TOTAL_MB` | `1024` | Keep total logs under N MB via oldest-first cleanup |
+
+## Error Handling
+
+| Error | Action |
+|-------|--------|
+| `.prizmkit/plans/feature-list.json` not found | Tell user to run `feature-planner` skill first |
+| JSON parsing failed | Use the bundled Python validation/status scripts instead of external JSON tools |
+| AI CLI not in PATH | Check Codex (`codex`), Claude (`claude`), or CodeBuddy (`cbc`) installation |
+| Pipeline already running | Show status, ask if user wants to stop and restart |
+| PID file stale (process dead) | `launch-feature-daemon.ps1` auto-cleans, retry start |
+| Launch failed (process died immediately) | Show last 20 lines of log: `Get-Content -Tail 20 .prizmkit/state/daemon/feature-daemon.log` |
+| Feature stuck/blocked | Use `reset-feature.ps1 <F-XXX> --clean --run` for a fresh retry |
+| All features blocked/failed | Show status, suggest daemon-safe recovery: `.\.prizmkit\dev-pipeline\reset-feature.ps1 <F-XXX> --clean --run .prizmkit/plans/feature-list.json` |
+| `playwright-cli` not installed | Browser verification skipped for playwright features (non-blocking). Suggest: `npm install -g @playwright/cli@latest; playwright-cli install --skills` |
+| `opencli` not installed | Browser verification skipped for opencli features (non-blocking). Install opencli for Chrome session-based browser verification |
+| PowerShell execution policy blocks script | Run `Set-ExecutionPolicy -Scope Process Bypass` for the current terminal |
+| Pipeline stop failed | Use `Stop-Process -Id <PID> -Force`; inspect AI CLI processes with `Get-Process \| Select-String codex` |
+| Deploy session failed | Pipeline completed but deploy session exited non-zero. Check `.prizmkit/state/features/deploy/<session_id>/logs/session.log`. Retry manually: `/prizmkit-deploy`. |
+| `.env.local` missing or incomplete | Warn: database connection variables not found. Suggest creating env file with required connection variables for the project's database |
+| Database unreachable | Warn: database features will produce mock-only tests. Suggest checking database service status and connection credentials |
+| Migrations not applied | Warn: tables or schema referenced in migration files not found in database. Suggest applying pending migrations |