agentsclimarketplace
Sign in

Review functionality

Skill gabriel-f-santos/coding-agents/skills/review-functionality

Review implemented code for conformance to a phase/feature contract and for correctness bugs. Use to check that what was built actually meets its acceptance criteria — "revisa se a feature faz o que devia", "conformance review", "review against the spec", "achou bug na fase?". Reads the phase contract (acceptance criteria, deliverables) + the diff and reports, confidence-based, whether each criterion is met and any correctness defects (logic, edge cases, error paths, concurrency). Read-only — reports findings, does not fix. Designed to run standalone or as a subagent of review-phase. Do not use for security (review-security) or style/cleanup (review-quality).From its SKILL.md

Install
npx -y skills add gabriel-f-santos/coding-agents --skill review-functionality

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.0 KB, 515 tokens by cl100k_base, as published. Nobody here has run it

review-functionality — does it do what the contract says?

Review the changed code against the phase contract, not in a vacuum. The differentiator vs a generic code review: every finding ties back to an acceptance criterion or a concrete correctness defect.

Inputs

  • The contract: docs/phases/phase-NN-<slug>.md — Objective, each SI's acceptance criteria and deliverables. (As a subagent, you receive the relevant excerpt.)
  • The change set: the files/diff to review (git diff).

What to check

1. Conformance (against the contract)

  • For each acceptance criterion: is it actually implemented and satisfied by the changed code? Quote the criterion; cite the code that meets it — or flag it unmet/partial.
  • Are all deliverables present (files, endpoints, migrations, tests the SI promised)?
  • Scope drift: anything implemented that the contract didn't ask for (note it), or a criterion silently skipped.

2. Correctness (defects in the changed code)

  • Logic: off-by-one, inverted conditions, wrong operator, incorrect default.
  • Edge cases: empty/null/huge inputs, boundary values, unicode, timezones, money rounding.
  • Error paths: unhandled failure, swallowed exceptions, partial writes without rollback, missing transaction around multi-step mutations.
  • Concurrency/idempotency: races, double-submit, retried-but-not-idempotent operations, lost updates.
  • Contracts: does the change break a caller, an API/wire format, or a DB constraint (uniqueness, length) reachable with legitimate input?
  • State & side effects: related entities updated/cascaded correctly; derived values can't collide or overflow.

Confidence & output

Report HIGH only (clear issue confirmed in the changed code); MEDIUM as "needs verification"; drop LOW. For each finding return: {dimension: functionality, severity, confidence, file:line, acceptance_criterion (quoted, if conformance), issue, evidence, why_it_matters, fix}.

Lead with failed/partial acceptance criteria (these are the highest-value findings), then correctness defects. If everything conforms and you find no defects, say so plainly — don't invent findings.

What ships with it: 1 file

222 B alongside SKILL.md

agents/

Gives 0 of the 12 instructions most review quality skills give in 515 tokens

Counted across 1,048 of the 1,783 authors here whose files we hold, read 2026-08-07

  • Ask questions one at a timein 81 of 1048, across 64 files
  • Provide a recommended answer for each questionin 73 of 1048, across 50 files
  • Explore the codebase instead of asking answerable questionsin 66 of 1048, across 42 files
  • Resolve dependencies between decisions one-by-onein 42 of 1048, across 17 files
  • Interview the user relentlessly about the planin 38 of 1048, across 13 files
  • Order findings by severityin 31 of 1048
  • Resolve each branch of the decision treein 27 of 1048, across 5 files
  • Run a grilling sessionin 26 of 1048, across 5 files
  • Update CONTEXT.md immediately when a term is resolvedin 26 of 1048, across 11 files
  • Propose precise canonical terms for vague languagein 25 of 1048, across 7 files
  • Create documentation files lazilyin 24 of 1048, across 5 files
  • Assign severity to every findingin 24 of 1048

Said here and by no other author read

  • verify all promised deliverables are present
  • flag unmet or partially met acceptance criteria
  • note scope drift or skipped criteria
  • check for logic errors in the changed code
  • check for unhandled errors and missing transaction rollbacks
  • check for concurrency and idempotency defects

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,149. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.