Review Tasks

SkillFiles & storage

Reviews an implementation task plan against its approved proposal, specifications, design, and test plan, finding only defects that can be corrected in the task files; use when asked to review tasks, review an implementation plan, verify task and automated-test coverage, or check whether a task breakdown is ready for autonomous implementation.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Review Tasks skill

What this skill tells your AI

The instructions your AI receives, as published by codagent-ai/agent-skills in skills/review-tasks/SKILL.md and read by ahel’s review.

Review whether the task index and detailed task files are a faithful, complete, executable translation of the approved proposal, specifications, design, and test plan.

Read all definition and task artifacts before judging the plan. Inspect relevant repository structure when needed to verify feasibility, paths, dependencies, test placement, or implementation boundaries. The definition artifacts are immutable inputs, not review targets.

Review focus

Check that the tasks:

  • cover approved requirements, scenarios, design obligations, migration, deliverables, and assigned integration or E2E tests without adding scope;
  • are self-contained for implementors with no hidden session context, including exact artifact citations, constraints, acceptance criteria, and done conditions;
  • form meaningful, right-sized delivery units with workable dependencies and ordering;
  • keep tests, ordinary documentation, setup, and refactoring with the behavior they support;
  • identify relevant implementation surfaces without prescribing line-by-line code; and
  • contain no duplicate, contradictory, orphaned, or unreachable work.

Do not ask for new product, scope, architecture, or testing-strategy decisions. If a concern requires changing an approved artifact, it is outside this review and must not become a task finding. Do not review implementation correctness or modify files.

Report

Every finding must be correctable only in the task index or detailed task files. Cite the exact task section and controlling artifact section, explain the implementation risk, and recommend the specific task edit. Rank findings by impact and end with a readiness assessment; say explicitly when no actionable task defect remains.

## Findings

### [high|medium|low] <title>
- Task: <task file and section>
- Controlling artifact: <artifact and section>
- Risk: <implementation consequence>
- Correction: <specific task edit>

## Readiness Assessment
<Whether the task plan is ready and why.>

Signals

GitHub stars
30
Forks
1
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
review-tasks
Source
github.com/codagent-ai/agent-skills