Skip to main content

Implement Requirement

In brief: Implement Requirement ships product behaviour for a user story or scenario under implement.policy.md, with IMPLEMENTED traceability and lifecycle ready by default.

When to use​

When you have a story or scenario id and want the agent to implement the product change (Development), not author tests or run the full Run QA composite. In the platform Workflows list, this workflow lives under Development (between Requirement Planning and Test Automation)—same grouping as this docs sidebar.

Why implement (policy traceability)​

Ad-hoc “implement this story” prompts in a coding agent leave no durable link between how the work was done and which requirement it satisfied. The implement workflow fixes that by running under implement.policy.md and reporting through TestChimp’s agent-action ledger:

  • Each run uses an explicit workflow_execution_id, workflow_id: implement, and the resolved policy_file / policy_version on every mutative step.
  • Finished work is recorded as IMPLEMENTED on the user story and scenario ordinals, with lifecycle status moved to ready by default (policy-configurable).

That matters when something goes wrong later. If an issue is opened against a story (regression, missed acceptance criterion, bad edge case), you can open the story’s Activity tab and see which policy version guided the implementation—not a one-off chat transcript. You then iterate implement.policy.md (coding standards, out-of-scope rules, self-review bar, when to run quality checks) and bump version so the next implement run is auditable again.

This is the development-side mirror of composite QA traceability in Workflows intro — Traceability: policies are how teams turn agent behaviour into something you can improve after production feedback, including via Fix issue when the failure is tracked as an issue id.

Prompt​

/testchimp implement US-181
/testchimp implement scenario 2205
/testchimp implement 42

Scope​

Explicit identity required — story or scenario ordinal (US-n / TS-n or bare number). No automatic branch/last-run scoping. Related scenarios are discovered by grepping plan scenario frontmatter for the parent story id.

Policy​

Loads plans/knowledge/policies/implement.policy.md for coding standards, out-of-scope rules, and Post-implement lifecycle status (default ready; set skip to leave status unchanged).

What it does​

  1. Analyze — Load the story/scenario from the platform; find related scenarios; note gaps/ambiguities (optional DeFOSPAM via Run Requirement Quality Checks when the agent judges it useful).
  2. Plan — Concrete code deltas covering in-scope scenarios; wait for approval.
  3. Execute — Implement, then self-review (acceptance gaps, logic, performance, smells).
  4. Report — report-agent-action with IMPLEMENTED per finished story/scenario; update-plan-items-lifecycle-status to ready (unless policy overrides); close with ACTION_COMPLETED.

See also​