How to use the Archify Review agent skill

What is an agent skill and what does archify-review make the agent do?

An agent skill is a folder containing a SKILL.md file with instructions that an AI agent loads when a task matches the skill's description. The archify-review skill is one such file. Its description says it reviews Archify issues, pull requests, or code through value, cost, and impact, and it is meant for issue triage, change reviews, and code quality assessments.

The skill is a review method, not a tool that runs checks on its own. It guides the agent's reasoning along three dimensions. Value covers the real problem, who benefits, how often it occurs, its consequences, and whether it is worth addressing now. Cost covers implementation and verification effort and the future cost of understanding and maintaining the code, and it favors existing capabilities and proportional solutions. Impact covers affected behavior and modules, consequences for users and future development, and whether the change improves maintainability or adds coupling.

• Checks facts against the target revision and separates verified findings from assumptions.
• Points out simpler or better alternatives and lists open questions that matter to the decision.
• Considers maintainability, complexity, compatibility and performance where they are relevant to the problem.

Install the archify-review skill from the ZIP or the Claude apps

Download the ZIP for the skill from the Archify page on this directory. Unzip it so that the archify-review folder, containing SKILL.md, lands in one of two places. For a personal setup, use ~/.claude/skills/. For a single project, use the .claude/skills/ folder inside that project.

If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills. The source for this skill lists only one file, SKILL.md, so the folder you install should contain that file at minimum.

• Personal install: ~/.claude/skills/archify-review/
• Project install: .claude/skills/archify-review/ inside your project folder
• Claude apps: Settings → Capabilities → Skills, then upload the skill

Trigger the skill with natural requests

You do not need a special command. The agent decides whether to use a skill by matching your request against the skill's description, which mentions Archify issues, pull requests, code, issue triage, change reviews and code quality. Say what you want reviewed and mention Archify, so the match is clear.

Be specific about the target. Give the issue number, PR link, or file path, and say what decision you need to make. The skill works best when it has a concrete revision to check against.

• "Review this Archify issue and tell me if the problem is worth solving."
• "Assess this PR for value, cost and impact before I decide whether to merge."
• "Do a code quality review of the parser module in Archify and list open questions."

Example reviews you can ask for

Issue triage: ask the agent to judge whether an incoming report describes a real problem, how often it occurs, and whether it deserves attention now. The output should give reasons and point to what would resolve the remaining uncertainty.

Change review: ask whether a proposed change uses existing capabilities, what it touches, and whether it adds coupling or constraints. The skill asks the agent to explain the approach's cost and impact and to suggest better alternatives if it sees them.

Code quality assessment: ask for a value, cost and impact read on a piece of code, with verified findings kept separate from assumptions.

What the skill needs and how it behaves

The skill lists no requirements. It does not run scripts and does not use the network, according to its listing. Its guidance points to the repository's REVIEWING.md and CONTRIBUTING.md files for project-specific rules, and the agent consults them as needed. Those files come from the Archify repository, so the agent can only use them if it has access to that code.

Because the skill is marked internal, its guidance is written for Archify. Its review questions are general enough to apply to other projects, but the repository-specific rules it cites are not.

• Requires: nothing listed
• Runs scripts: no
• Uses network: no

Safety notes and limits

The skill does not execute code or send data anywhere, based on its listing. It still shapes what an agent concludes, so treat its output as an evidence-based opinion rather than a verdict. The agent may be wrong or incomplete, especially if it cannot check the target revision.

The skill is not a replacement for a human maintainer's final merge decision. Read the cited evidence yourself before acting on a recommendation, and confirm that the agent looked at the exact revision you are reviewing.

• Do not use it on projects other than Archify without checking the output, since it points to Archify-specific files.
• Do not treat the result as a merge approval.
• Ask the agent which checks it ran and which claims are assumptions.

Tips for better reviews

Give the agent the revision, issue, or PR you mean, and ask it to verify facts against that revision. Ask it to label each point as verified, based on existing evidence, or assumed. This is the behavior the skill is designed to produce, so asking for it explicitly helps.

If a review feels generic, ask the agent to name the specific module or behavior affected and to explain why each concern matters. The skill asks for detail scaled to the importance of each concern, so you can also tell it to focus on one dimension.

• Ask for open questions that matter to the decision, not a long list of minor points.
• Ask for simpler alternatives explicitly, since the skill includes them when it finds them.
• Compare the agent's findings with the project's own REVIEWING.md before deciding.