process-feedback

Resolves one feedback cluster against primary sources and returns a canonical verdict.

Updated Mar 21, 2026
One-click install
npx skills add https://github.com/mikestopcontinues/agent-skills --skill process-feedback-mikestopcontinues
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: process-feedback
Source: https://github.com/mikestopcontinues/agent-skills/tree/main/claude/skills/process-feedback
Command: npx skills add https://github.com/mikestopcontinues/agent-skills --skill process-feedback-mikestopcontinues

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Review feedback often arrives as scattered findings across multiple review files, and someone must investigate each cluster against primary sources before a decision can be made. This Skill handles exactly one feedback cluster end-to-end, deep-diving ambiguous cases against repository state, git history, and vendor documentation, then returning a single structured verdict. ## Core Features & Use Cases - Per-cluster resolution: Reads one cluster, its artifact, and the referenced rNN- review files, then investigates each member finding's exact claim and location. - Two canonical verdict shapes: Emits either a verified-resolution (with mandatory citation and patch suggestion) or a narrowed-options-brief (with per-option trade-offs and a recommendation), never a third shape. - Decision-shaped hand-back: Detects clusters that should route to lock-decisions and returns a hand-back marker instead of a verdict. - Use Case: A triage step routes an ambiguous cluster about API versioning to this handler; it fetches the matching vendor docs, confirms the correct approach, and returns a verified-resolution with a citation and concrete patch suggestion. ## Quick Start Process feedback cluster C3 about authentication token expiry using the artifact at docs/auth-spec.md and review files r01-security.md and r02-architecture.md.

Frequently Asked Questions about process-feedback

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I process a triaged feedback cluster from a document review?▼

Read the cluster description, the artifact it references, and every rNN- review file containing member findings. Then deep-dive ambiguous clusters against primary sources and return exactly one verdict: verified-resolution or narrowed-options-brief.

What is the difference between verified-resolution and narrowed-options-brief?▼

A verified-resolution means a primary source definitively settles the cluster and requires a citation plus a patch suggestion. A narrowed-options-brief is used when no source resolves it, presenting at least two options with trade-offs and a recommendation for the captain to decide.

When should a feedback cluster be handed back instead of resolved?▼

Hand back when the cluster is decision-shaped: its route is lock-decisions, members come from r*-scope.md files, or members are major/blocker findings from r*-architecture.md. Emit the hand-back marker and do not produce a verdict.

Can I use web search to resolve ambiguous review findings?▼

Yes, but WebFetch is limited to vendor docs and RFCs whose version matches the project's declared version, and WebSearch is only for discovering which doc URL to fetch. Repository reads and git commands should be exhausted first.

What happens if the deep-dive finds nothing conclusive?▼

Return a narrowed-options-brief with options like accepting the current artifact state or gathering more info via a specific action. Never invent a third outcome header such as research-pending, since the lifecycle skill only pattern-matches the two canonical headers.