ahepi/deepreason/review-response
Claude Code and agent skill by ahepi/deepreason
- community
Checking that the SKILL.md file is still in the repository.
The find -> verify -> refute-or-accept-in-writing -> act loop for every external review received (FR-21, FR-25), and the author defect ledger that reviews are required to cite. This loop replaced the unattended driver; see MODULES.md.
Install
npx vskill@latest install ahepi/deepreason/review-responseSKILL.md
Review Response (new in 0.5; the loop that replaced M2)
<!-- PROMPT-CORE-BEGIN --> A review is evidence, not a verdict. Every review received gets a written disposition; every finding in it gets exactly one of two fates.
- VERIFY FIRST: re-establish each finding against the sources or by
running code (FR-25) before acting. A reviewer without execution access returning MISSING on an execution-dependent claim is correct behavior - the verification is yours to run, not theirs to guess.
- REFUTE OR ACCEPT, in writing, per finding:
- refuted: state the evidence, in the disposition, with a test
pinning the refutation where possible. A refuted finding is as valuable as an accepted one - record it, never silently drop it;
- accepted: name the action taken in the same disposition. An
accepted finding without an action is not yet accepted. A reviewer's REASONING can be wrong while its WARNING is right; dispose of the two separately.
- THE DISPOSITION is one document per review: verdict table, narrow
green (what the reviewer's packet did and did not contain), and the per-finding fates. It cross-links from the artifact reviewed.
- AUTHOR DEFECT LEDGER: maintain a running list of the author's own
verified failure modes (one line each: defect, instance, direction). Feed it to adversarial reviewers - naming the author's documented bias measurably sharpens attack. Two integrity rules: entries only for VERIFIED defects, and errors in BOTH directions recorded - a ledger that only shows one direction is itself the tidy story it warns against.
- ESCALATION DISCIPLINE (FR-21): a review that shows you closed a
question that was the owner's to decide is accepted by REOPENING the question - restating options, moving your view to a marked view - never by defending the closure. Having been corrected for closing is not a license to feign neutrality: state views as views.
- The loop ends when every finding has a fate and every action is
done or filed. Then, and only then, the next review round. <!-- PROMPT-CORE-END -->
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/review-response/SKILL.md
Popularity
- Repository stars
- 141
- Repository forks
- 15
- 7-day trend score
- 0
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/ahepi/deepreason/review-response)