
Goldfish Review Resolver
Triages reviewer findings on a document and proposes or applies fixes to it
- Repository
- google/elephant-goldfish-model
- Stars
- 9,965
- Licence
- Apache-2.0
- Files
- 2 files
What it does
This skill lets an agent take a list of problems flagged by reviewers of a document (the Goldfish review format used in the Elephant Goldfish model project) and sort each one into a category: contest it with evidence, escalate it to the user, propose a text fix, or decline it as a minor nit. It writes a markdown report of these decisions, then pauses and waits for you to choose: auto-accept the proposed fixes, skip them, or walk through each item one by one. Only when you accept does it edit the target document.
Great for
- Authors who received structured review findings on a design or technical document and want a first-pass triage
- Teams running the Elephant/Goldfish multi-reviewer workflow who want consistent resolution of flagged items
Skip it if
- General proofreading or editing of a document with no reviewer findings; the skill expects a specific findings list format
- Auto-accepting fixes on important documents without reading the report, since the agent may misjudge a finding and edit the file directly
Try asking
- “Resolve the findings from my Goldfish review of design/spec.md lines 40-120 and write the report”
- “Walk me through each proposed fix interactively before changing anything”
- “Continue to the next phase without applying any fixes, just give me the report”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 2 original files, the Apache-2.0 licence and an attribution note.
egm-resolve.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip egm-resolve.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip egm-resolve.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Run it in the same conversation session that produced the document so the agent can recall why choices were made, as the skill recommends.
- Read goldfish_review_report.md before replying 'auto accept', especially for ESCALATE and CONTEST items, which are excluded from auto-accept but need your judgement.
- Use 'interactive breakdown' for high-stakes documents so each proposed edit is explained before it is applied.
What's inside
- agents/openai.yaml
- SKILL.md
Redistributed unmodified from google/elephant-goldfish-model under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-11-02. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.