More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Post-Review: The Second Reading Before You Call It Done

    1

    Run a systematic post-implementation review before marking work done — six lenses, graded findings, and a fix-and-verify loop that ends clean, not merely acknowledged.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Post-Review: The Second Reading Before You Call It Done

    Post-Review: The Second Reading Before You Call It Done

    Example session with this skill installed

    A PR adds 'export CSV' to our admin dashboard. Review it before merge — spec, build, inputs, fit, edges, standards — and grade what you find.

    • Read your context and instructions
    • Compiled the post-review

    Lens 1, spec: the spec said 'export filtered results' — the code exports all results, ignoring the active filter. Blocker. Lens 2, build: type checker passes in strict mode, linter passes, production build succeeds, no console errors, no debug leftovers. Clean. Lens 3, inputs: no validation on the row limit — a 2M-row export would OOM the server. Blocker: add a cap and stream the response. Error messages were also checked: they say what failed and what to do next. Lens 4, fit: the code follows the existing patterns for similar exports, no duplicated logic, no new dependency. Clean. Lens 5, edges: an empty result set produces a 0-byte download with no message. Should-fix: return a header-only file plus a notice. Concurrency and scale were checked — the streaming response handles both. Lens 6, standards: the endpoint lacks the admin-role check present on every other admin route. Blocker. Three blockers on a 'simple' feature — that is the review earning its keep. Fix loop: the filter is respected, the row cap and streaming are in, the role check is added, the empty-state notice ships. Build and affected checks re-run, each blocker confirmed closed by evidence, not by a comment. Verdict: APPROVED after re-review.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    'Done' means reviewed. Code that has not been read critically by a second pair of eyes — even your own, a day later — ships with the author's blind spots intact. This skill is that second reading, structured so nothing important gets skimmed: six lenses read the diff six times (spec conformance, build health, input handling, architecture fit, edge cases, standards), every finding is graded blocker, should-fix, or nit, and the fix loop ends when the six lenses come back clean — not when the findings list is merely acknowledged.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse All Claude Skills.

    A review finds what the lenses cover; a clean review is not proof of zero bugs — it is proof the six lenses came back clean.
    Nits batch: they do not block a merge, but they are not dismissed either — batch them, do not forget them.
    Lightweight reviews apply to dependency adds and config changes; the full six-lens pass is for features, fixes, and refactors.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Permissions

    No special permissions declared or detected

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    Before. Never mark a feature done without it. The review is part of the work, not a ceremony after it.

    A blocker is wrong behavior, a security hole, or a broken build — fix before merge. A should-fix is a real quality issue: fix now, or file a ticket with an owner. 'Later' without a ticket means never.

    Lightweight: a security and license glance. The full six-lens pass is for features, bug fixes, refactors, and core utility changes.