Skip to content
Skills Library
AI & agents · Building & coding

implementing

/implementing

Performs Task Contract-scoped implementation work and records Implementation Evidence. Use when a coordinator hands off implementation phase work, code or document outputs must be changed inside an accepted Task Contract, and changed outputs, self checks, limits, or contract deltas must be returned. Do not use for Verification Evidence, Review Evidence, Challenger Evidence, Task Evidence, User Judgment, Mission baseline drafting, or Mission closure.

Repository
geas ↗
Licence
Apache-2.0
Last repository change
19.05.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
2
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 19.05.2026, without references to retired models or APIs.

Licence

Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.

What it does and when to use it

Performs Task Contract-scoped implementation work and records Implementation Evidence. Use when a coordinator hands off implementation phase work, code or document outputs must be changed inside an accepted Task Contract, and changed outputs, self checks, limits, or contract deltas must be returned. Do not use for Verification Evidence, Review Evidence, Challenger Evidence, Task Evidence, User Judgment, Mission baseline drafting, or Mission closure.

Original author description

Performs Task Contract-scoped implementation work and records Implementation Evidence. Use when a coordinator hands off implementation phase work, code or document outputs must be changed inside an accepted Task Contract, and changed outputs, self checks, limits, or contract deltas must be returned. Do not use for Verification Evidence, Review Evidence, Challenger Evidence, Task Evidence, User Judgment, Mission baseline drafting, or Mission closure.

How it works

  1. Job
  2. Operating Stance
  3. Inputs
  4. Workflow
  5. Resources
  6. Gotchas
  7. Stop Conditions
Original SKILL.md headings · 116 lines, 883 words
SKILL.md preview
---
name: implementing
description: Performs Task Contract-scoped implementation work and records Implementation Evidence. Use when a coordinator hands off implementation phase work, code or document outputs must be changed inside an accepted Task Contract, and changed outputs, self checks, limits, or contract deltas must be returned. Do not use for Verification Evidence, Review Evidence, Challenger Evidence, Task Evidence, User Judgment, Mission baseline drafting, or Mission closure.
---

# Implementing

## Job

Perform the requested implementation inside the current Task Contract, preserve the implementation context as Implementation Evidence, and return the Evidence ref and limits to the caller.

`implementing` owns implementation changes and Implementation Evidence only. It does not decide Task acceptance, write Task Evidence, verify completion for the User, or review its own work as a separate review role.

## Operating Stance

Act as a contract-bound implementer.

- Read first, then plan, then change.
- Prefer the smallest change that satisfies the Task Contract.
- Keep changes reviewable for the next verifier and reviewer.
- Treat self checks as handoff hygiene, not verification or acceptance.
- Surface scope gaps, environment gaps, and contract deltas instead of expanding authority.
- Leave clear implementation context: what changed, why it changed, what was checked, and what remains limited.

## Inputs

Required inputs:

- Accepted Mission Spec ref.
- Accepted Mission Design ref.
- Current Task Contract ref.
- Task id.
- `read_first` refs for the Mission baseline, Task Contract, target outputs, and relevant prior Evidence.
- Handoff focus with the requested implementation scope.
- Target files, artifacts, or output refs.

Optional inputs:

- Environment context from the coordinator.
- Prior Implementation, Verification, Review, Challenger, or Task Evidence refs.
- User constraints that are already part of the accepted Task Contract or handoff packet.

## Workflow

### Normal

1. Read every `read_first` ref before changing outputs.
2. Identify the Task Contract scope, target outputs, acceptance criteria, verification expectations, environment context, and handoff focus.
3. Check stop conditions before implementation.
4. Create a short implementation plan before changing outputs.
5. Make only the changes covered by the Task Contract and handoff focus.
6. Run implementation-side self checks that are available inside the agreed environment.
7. If self checks reveal an in-contract fix, fix it and repeat the relevant self check.
8. If self checks still fail after available in-contract fixes, stop and report the unresolved failure.
9. If work reveals a blocking contract delta, scope gap, missing environment dependency, or required User decision, stop and report it instead of broadening scope.
10. If work reveals a non-blocking contract delta, keep it explicit in Implementation Evidence and caller output.
11. Prepare Implementation Evidence with `references/implementation-evidence.md`.
12. Ask the `geas-cli` adapter to record Implementation Evidence.
13. Return the Evidence ref, changed outputs, self checks, limits, and contract deltas to the caller.

### Implementation Plan

Before editing, write a concise plan that names:

- Target outputs to change.
- Ordered implementation steps.
- Self checks to run before handoff.
- Contract boundary or environment condition that would stop the work.

If the plan reveals missing scope, missing target refs, unavailable environment support, or a required User decision, stop before changing outputs.

### Record Failure

If `geas-cli` cannot record the Evidence, return the prepared payload and the record failure. Do not replace the missing Evidence ref with an informal completion claim.

### Contract Delta

Classify each contract delta before deciding whether to continue:

- What the contract currently permits.
- What new work, file, dependency, behavior, or decision is needed.
- Whether implementation can continue safely without that delta.
- Whether any changed outputs already exist and how to inspect or revert them.

A blocking delta requires a stop report. Blocking deltas include required work outside the Task Contract, a new User decision, a new dependency or configuration change not covered by the contract, or any change that would alter accepted behavior, scope, acceptance criteria, or verification expectations.

A non-blocking delta can be recorded in Implementation Evidence. Non-blocking deltas include follow-up contract refinement candidates, reflection candidates, or implementation notes that do not require broader authority before the current handoff.

## Resources

| Resource | Use When |
| --- | --- |
| `references/implementation-evidence.md` | Writing Implementation Evidence, checking record readiness, and reporting implementation stop reasons. |

## Gotchas

- Do not change outputs outside the Task Contract or handoff focus.
- Do not start changing outputs before forming the implementation plan.
- Do not use self checks as Verification Evidence.
- Do not write Verification Evidence, Review Evidence, Challenger Evidence, Task Evidence, or User Judgment.
- Do not silently install tools, add connectors, start services, or change project configuration beyond the environment choices supplied by the coordinator.
- Do not make repo dependency or config changes for environment setup unless the Task Contract covers that setup.
- Do not continue when a required `read_first` ref cannot be read.
- Do not hide contract deltas inside implementation notes; classify them as blocking or non-blocking.
- Do not record Implementation Evidence before changed outputs, self checks, limits, and contract deltas are summarized.

## Stop Conditions

Stop and return a handoff failure report when:

- A required `read_first` ref is missing or unreadable.
- The Task Contract, task id, handoff focus, or target output refs are missing.
- The next change is outside the Task Contract or handoff focus.
- A required toolchain, service, secret, connector, or runtime environment is unavailable.
- Self checks still fail after available in-contract fixes.
- Self checks fail and the fix is outside the Task Contract.
- A blocking contract delta is found.
- A User decision is required before continuing.
- The next step belongs to verification, review, challenge, Task acceptance, or Mission closure.
Files in this package (2)
  • SKILL.md 6 KB
  • references/implementation-evidence.md 4 KB

Versions

RepositoryScoreLicence
This page · choam2426/geas90Apache-2.0
fynnhaupt/skills90MIT
azat-io/ai-config90MIT

Database snapshot: 13.11.2026 · package version 5a165102bbfb (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 90verifying
    Performs Task Contract-scoped verification of changed outputs and records Verification Evidence. Use when a coordinator hands off verification phase work with a current Task Contract, Implementation Evidence, target outputs, and verification environment; run available checks,…
  • 90reviewing
    Performs Task Contract-scoped review of changed outputs and role Evidence, then records Review Evidence. Use when a coordinator hands off review phase work with a current Task Contract, Implementation Evidence, Verification Evidence, target outputs, and review focus; inspect…
  • 95mission-control
    Run ONE outer-loop iteration of a long-running mission (default: the V1 mission) — observe mission state, pick the top backlog item, route it through design-doc-creator → sprint-planner → sprint-executor → sprint-evaluator with the mission''s model routing policy, record a log…
  • 100he-work
    Executes bounded implementation work from approved specs, plans, issues, or small fixes by editing files, running validation, preserving unrelated work, and recording rollback and handoff evidence. Use when asked to implement, execute a plan, apply changes, or build a scoped feature.
  • 100task-writing
    Use when writing an implementation plan for a multi-step job, or when preparing a task text for an agent (scout, reviewer, maintainer); especially if parallel work is to be given to more than one agent. The main chat (central agent) loads it.
  • 90mission-ledger
    Magentic-style Task/Progress Ledger orchestration for /blueprint, /forge, /cruise, and /mission. Use whenever planning or executing multi-step work on the board.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.