From prfaq-dev
Runs simulated PR/FAQ review meeting with four personas (engineer, customer, executive, builder) debating weak spots in prfaq.tex, producing decisions and summary file.
How this command is triggered — by the user, by Claude, or both
Slash command
/prfaq-dev:meeting [path/to/prfaq.tex]This command is limited to the following tools:
The summary Claude sees in its command listing — used to decide when to auto-load this command
# PR/FAQ Review Meeting
Run an interactive review meeting where four personas — a principal engineer, a target customer, a skeptical executive, and a builder-visionary — debate the weak spots in your PR/FAQ document. You are the PM and final decision-maker.
## Steps
1. **Find the document.** If `$ARGUMENTS` specifies a path, use it. Otherwise, search for `prfaq.tex` in the project root using Glob. If no document exists, tell the user to run `/prfaq` first — the meeting validates an existing document, it doesn't generate one.
2. **Read the meeting guide.** Load `${CLAUDE_PLUGIN_ROOT}/ski...Run an interactive review meeting where four personas — a principal engineer, a target customer, a skeptical executive, and a builder-visionary — debate the weak spots in your PR/FAQ document. You are the PM and final decision-maker.
Find the document. If $ARGUMENTS specifies a path, use it. Otherwise, search for prfaq.tex in the project root using Glob. If no document exists, tell the user to run /prfaq first — the meeting validates an existing document, it doesn't generate one.
Read the meeting guide. Load ${CLAUDE_PLUGIN_ROOT}/skills/prfaq/references/meeting-guide.md for the full meeting flow, synthesis guidelines, and persona details.
Run the pre-meeting scan. Read the full .tex document. Extract \prfaqstage{value} to calibrate expectations (see Stage Calibration in the meeting guide). Identify 5-8 hot spots: sections with unsupported claims, vague language, thin evidence, risk rating mismatches, or gaps between press release confidence and FAQ hedging. Rank each as Critical, Warning, or Suggestion — calibrated to the document's stage.
Present the agenda. Show the user the hot spots ranked by severity and offer scope options via AskUserQuestion:
Run the debate loop. For each selected agenda item: a. Quote the specific text under discussion b. Launch all four persona agents in parallel using the Task tool:
subagent_type: "prfaq:meeting-engineer" (Wei)subagent_type: "prfaq:meeting-customer" (Priya)subagent_type: "prfaq:meeting-executive" (Alex)subagent_type: "prfaq:meeting-builder" (Dana)
c. Synthesize their responses into a dramatic debate narrative (not concatenation)
d. Present the decision via AskUserQuestion: Revise / Keep as-is / Research / Defer
e. Show cascade consequences (which other sections are affected)Present the post-meeting summary. List all decisions, build a revision queue with specific feedback directives for /prfaq:feedback, and note deferred items.
Persist the summary. Write the meeting summary to ./meetings/meeting-summary-YYYY-MM-DD.md (today's date). If that file already exists, append a counter (-2, -3, etc.). See Phase 3b in the meeting guide for the full file format. Tell the user where the file was saved.
Migration: Before writing, use Glob to check for meeting-summary-*.md and meeting-hive-summary-*.md in the project root (same directory as the .tex file). If any are found, move them to ./meetings/ using the Read and Write tools (read content, write to new path, delete old file via Bash rm). Tell the user: "Moved N meeting summary file(s) to ./meetings/ for organization."
Offer to apply revisions. If the revision queue is non-empty, tell the user to run /prfaq:feedback (no arguments) to automatically discover this meeting summary and apply all directives sequentially.
npx claudepluginhub punt-labs/claude-plugins --plugin prfaq/refineGenerates a PR/FAQ document from prior customer research artifacts, following Amazon's Working Backwards process. Saves output as markdown and prompts for next phase.
/refineAnalyzes existing PRD for gaps using a section checklist, generates markdown gap analysis, and refines via targeted interactive questions.
/writing-reviewSpawns parallel creative writing reviewers (editorial, prose quality, dialogue) to analyze a draft and synthesizes their feedback into a unified editorial report.
/reviewReviews documents or free-text topics with multiple AI models independently, then synthesizes and converges on a unified review.
/grillStress-tests a plan, design, or PRD by interrogating decisions one at a time, grounding each question in the codebase to surface contradictions and assumptions before implementation.
/review-design-docInteractively reviews design doc sections on key decisions, risks/mitigations, and open questions; explains each, solicits feedback, and applies updates.