agentsclimarketplace

He code review

Skill jscraik/Agent-Skills/Plugins/harness-engineering/fixtures/budget-archive/2026-04-21/deferred-store/skills/code_quality_review/he-code-review

Governed skill foundry and Skills SDK for Codex/AI coding agents: author, validate, evaluate, and sync runtime projections through ask.

Install
npx -y skills add jscraik/Agent-Skills --skill he-code-review

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

One thing to look at

  • 8 stars8 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.

What its author says it does

Copied from the file, not written here

Review PRs, branches, diffs, and workflow artifacts for package-level go/no-go readiness with severity-ranked synthesis. Use when users need readiness synthesis rather than detailed technical-risk critique.

SKILL.md

4.5 KB, as published. Nobody here has run it

he-code-review Entry

Use when package-level readiness, merge risk, release go/no-go, or tracked PR closure is the core question.

Context disposition: Apply the context-disposition policy: move important still-valid context to references and index it when meaningful; intentionally discard stale, duplicated, unsafe, superseded, or low-signal text.

Philosophy

  • Prefer concrete blockers over broad commentary.
  • Treat readiness as an evidence problem, not a title or branch-name problem.

When to use

Use he-code-review when the user wants a package-level readiness verdict for a PR, branch, diff, or delivery slice.

Required inputs

  • PR, branch, diff, commit range, or artifact path under review.
  • Intended base branch, linked Linear issue, spec or plan paths, and acceptance IDs when tracked work is involved.
  • Available CI, reviewer-thread, and validation evidence.

Deliverables

  • Severity-ranked readiness findings with exact file or artifact evidence.
  • A go, go-with-conditions, or no-go recommendation.
  • Clear follow-up routing to he-work, autofix, security-ops, or GitHub PR workflow when required.

Core Contract

  • Resolve target, base, mode, and plan/spec links before judging.
  • Build an evidence pack from diff/base, changed files, checks, validations, review threads, Linear issue, branch/PR, spec, plan, and acceptance IDs.
  • Rank only concrete readiness issues as P0-P3; discard style-only, speculative, duplicate, and protected-artifact cleanup notes.
  • Emit go, go-with-conditions, or no-go.
  • Keep broad review read-focused unless an explicit mode permits safe autofix.

Procedure

  1. Resolve target, base, mode, Linear issue, spec, plan, and acceptance IDs.
  2. Build the evidence pack from diff, checks, validation, and review threads.
  3. Rank concrete findings and emit the readiness verdict.

Linear Readiness

For tracked delivery, verify:

Linear issue -> spec/source acceptance IDs -> plan units -> PR evidence -> validation

Do not treat branch names, PR titles, or resolver claims as closure. A clean go requires behavior and validation evidence tied back to Linear and the governing artifacts.

Validation

  • Each finding needs severity, exact location, evidence, impact, confidence, and remediation.
  • Block go for unresolved P0/P1, actionable reviewer threads, relevant failing checks, stale conflicts, missing validation, or missing Linear/spec/plan/PR traceability.
  • For PR evidence artifacts, run python3 Infrastructure/scripts/validation-and-linting/he_linear_traceability_lint.py <artifact-path> before returning a go verdict.
  • Route broad security-sensitive decisions to security-ops or a security reviewer.
  • Stop at the first failed gate.

Constraints

  • Redact secrets, credentials, tokens, and sensitive data by default.
  • Do not claim readiness without repository evidence.
  • Keep broad reviews read-focused unless an explicit mode permits mutation.

Anti-patterns

  • Treating CI alone as merge readiness.
  • Approving tracked PRs without Linear/spec/plan/PR traceability.
  • Reporting vague findings without a concrete code or artifact path.

Examples

  • "Review this PR and verify it closes the linked Linear QA issue."
  • "Check this branch against origin/main and give the go/no-go call."

Failure mode

If the base, target, review scope, or required tracker/spec/plan evidence is missing, stop and request that source instead of giving a false readiness verdict.

Gotchas

  • Do not treat passing CI alone as merge readiness.
  • Do not resolve reviewer threads from this skill unless explicitly in an autofix or PR-management mode.
  • Keep security-sensitive or policy-sensitive concerns routed to the appropriate specialist.

References

  • Full guide: Plugins/harness-engineering/fixtures/preserved-context/skills/he-code-review/SKILL.full.md
  • Review flow: Plugins/harness-engineering/fixtures/preserved-context/skills/he-code-review/references/codex-review-flow.md
  • Review modes: Plugins/harness-engineering/fixtures/preserved-context/skills/he-code-review/references/review-modes.md
  • Subagent routing: Plugins/harness-engineering/references/subagent-routing.md
  • Domain and QA routing: Plugins/harness-engineering/references/domain-model-routing.md, Plugins/harness-engineering/references/qa-intake-routing.md

Keep looking

Skills are one crate of 328,083. 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.