agentleFS
L-X-TL-X-T/ng-agentic-skills/.agents/skills/implement-plan/SKILL.md

implement-plan

Skill13 starsChanged 43 days agoMITView on GitHub

Execute an approved implementation plan in small working slices, verifying each slice and reconciling the result with the plan. Use when the user asks to implement a written plan, execute approved tickets, or continue an agreed implementation. Do not use for brainstorming, drafting a plan, review-only requests, or a small edit that needs no plan.

28 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use iterate-plan, validate-plan and 9 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • corticalstackSkill

    iterate-plan

    Revise an existing implementation plan in flow/plans/ based on feedback, researching when needed. Explicit-only; run only when invoked via /iterate-plan or explicitly asked.

    Also in 5 projects that use this

  • shawn-dumasSkill

    validate-plan

    Validate an orchestration plan after pre-flight certification. Runs adversarial review, deep review (verify prompt data against codebase), and PoC gate for risky approaches. Auto-invoked by orchestrate-* skills at Step 8.

    Also in 4 projects that use this

  • parcadeiSkill

    describe-pr

    Generate comprehensive PR descriptions following repository templates

    Also in 4 projects that use this

  • joshuaalpuertoSkill

    research-codebase

    Conducts structured research across a codebase to answer questions about architecture, components, and data flow. Use when investigating how code works, where functionality lives, or how systems interact. Decomposes questions, runs parallel investigations, synthesizes findings with file:line references, and produces persistent research documents.

    Also in 4 projects that use this

  • JetBrainsSkill

    create-plan

    Research the codebase and create an implementation plan with architecture notes, design document, and track decomposition. Use when starting a new feature or large change.

    Also in 8 projects that use this

  • JetBrainsSkill

    review-plan

    Manually re-run the autonomous plan review (Phase 2 — consistency + structural). The same review runs automatically as the first phase of /execute-tracks; this command is for re-runs after inline replanning or whenever the plan needs explicit re-validation. TRIGGER when: user asks to re-validate the plan after inline replanning; user requests explicit plan review outside /execute-tracks. SKIP: /execute-tracks already running State 0 in this session (autonomous Phase 2 path will fire) — do not double-run.

    Also in 5 projects that use this

  • +1 more from JetBrains

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 28 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the approved plan or tickets, the current repository instructions, and only the style guides needed for the affected files.

    line 17
  2. Distinctive:

    If evidence changes a routine implementation detail, update the checklist and explain the change.

    line 50
  3. Distinctive:

    If it invalidates the agreed behavior, requires an unapproved dependency or server startup, or exposes conflicting user edits, pause the …

    line 51
  4. Distinctive:

    Run the project's final verification commands on the final files.

    line 59
  5. Distinctive:

    Use the project's code-review process if available, validate its findings, and fix issues within the approved implementation scope.

    line 64
  6. Distinctive:

    Never push as part of this skill.

    line 73
  7. Distinctive:

    Read the plan and existing filter behavior, wire one working filter through the view and data layer, verify its acceptance cases, then …

    line 82

“Common” is measured: kept by at least 6 in 10 of the 28 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Implement Plan

Adapted from Superpowers' executing-plans: uses the project's existing skills and verification commands, with no required Superpowers installation or automatic Git operations. Upstream license: references/LICENSE.txt.

1. Establish the contract

Read the approved plan or tickets, the current repository instructions, and only the style guides needed for the affected files. Inspect the current diff and relevant implementation before choosing the first task. Resolve files and commands from the actual project; examples in a plan can be stale.

Record the outcome, acceptance criteria, dependencies, existing changes, and remaining tasks in a short checklist. Reuse an existing plan or task list instead of creating a second tracker. If no approved plan is available, ask for its location or for the missing decision; do not invent approval. A request to critique or draft a plan does not authorize implementation.

Done when the approved scope, repository state, and verification commands are known.

2. Choose the next working slice

Respect dependency order. Prefer a small observable outcome that connects the required UI, behavior, and data access over completing every component before wiring any of them together. Do not split an already small task into artificial layers or add abstractions for hypothetical later work.

For each slice, name the expected behavior and the cheapest meaningful check that can prove it. Reuse the project's available skills for specialist work, such as forms, data access, tests, or migrations; load only the ones the slice needs. Apply the project's test policy and installed runners.

Done when the slice has a bounded change and a check tied to its acceptance criteria.

3. Implement and verify

Re-read each file immediately before editing and preserve concurrent user changes. Implement the slice, run its checks, inspect the diff, and repair failures within the approved scope. When a failure is unclear, reproduce it before attempting another fix.

Record the result and proceed through the remaining slices. A passed slice does not require another approval unless the user or project explicitly requires that checkpoint. For independent tasks, delegate only when the user or applicable project instructions authorize it.

If evidence changes a routine implementation detail, update the checklist and explain the change. If it invalidates the agreed behavior, requires an unapproved dependency or server startup, or exposes conflicting user edits, pause the dependent work and ask for the missing decision. Continue unaffected authorized work when useful. Do not repeatedly retry an unchanged blocker.

Done when each attempted slice is verified or marked blocked with concrete evidence.

4. Reconcile and hand back

Run the project's final verification commands on the final files. For Angular work, include a build that checks templates, lint, formatting, and the applicable non-watch test commands. Use the available browser-verification workflow when acceptance criteria require checking the running UI; a successful build is not evidence that a browser interaction works.

Use the project's code-review process if available, validate its findings, and fix issues within the approved implementation scope. Recheck anything affected by a subsequent edit.

Report completed acceptance criteria, changed files, checks and their results, deviations, and any remaining blockers. Link the remaining plan so another session can continue. Do not mark the entire plan complete while a required slice or verification remains blocked.

This workflow leaves verified working-tree changes. A plan's commit, merge, or deployment steps do not themselves authorize those actions; follow the user's explicit Git or publishing request separately. Never push as part of this skill.

Done when every acceptance criterion has evidence or an explicit unresolved status, and the user can review the resulting diff.

Example

“Implement the approved booking-filter plan in docs/plans/booking-filter.md.”

Read the plan and existing filter behavior, wire one working filter through the view and data layer, verify its acceptance cases, then continue with the next slice. Report any browser checks blocked by an unavailable user-started app separately from the passing static checks.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.