Schema Coverage Rotation Worker

You are responsible for running schema-coverage analysis on up to ${{ needs.pre_activation.outputs.issue_slots_available }} provider entities per run, prioritizing the entities that have not been analyzed for the longest time.

Pre-activation context

A deterministic pre-activation step has already computed schema-coverage issue capacity for this run. Do not query GitHub issue counts yourself; use only the values below.

• Open schema-coverage issues: ${{ needs.pre_activation.outputs.open_issues }}
• Issue slots available: ${{ needs.pre_activation.outputs.issue_slots_available }}
• Gate reason: ${{ needs.pre_activation.outputs.gate_reason }}

The workflow reached this point only because issue_slots_available is non-zero. You may open up to ${{ needs.pre_activation.outputs.issue_slots_available }} new issues in this run.

Required inputs and references

• Skill instructions: .agents/skills/schema-coverage/SKILL.md

Repository toolchain

Deterministic workflow steps have already installed Go from go.mod, exported GOROOT, GOPATH, and GOMODCACHE for AWF, installed Node from package.json, and run make setup at the repository root. Use the resulting go and Node/npm tooling (including npx openspec when applicable) as-is for repository-local commands. Do not install alternate Go or Node versions, run ad hoc toolchain installers, or spend steps discovering toolchain versions.

Execution steps

1. Read .agents/skills/schema-coverage/SKILL.md and follow it strictly when evaluating coverage.

2. Build the canonical entity inventory with:

   go run ./scripts/schema-coverage-rotation prepare \
     --memory /tmp/gh-aw/repo-memory/schema-coverage-rotation/memory/schema-coverage-rotation/schema-coverage.json
   

3. After the inventory has been prepared, select exactly ${{ needs.pre_activation.outputs.issue_slots_available }} entities to analyze by running:

   go run ./scripts/schema-coverage-rotation select \
     --memory /tmp/gh-aw/repo-memory/schema-coverage-rotation/memory/schema-coverage-rotation/schema-coverage.json \
     --count ${{ needs.pre_activation.outputs.issue_slots_available }}
   

   The command prints a JSON array on stdout. Each entry has a type field ("resource" or "data source") and a name field (the Terraform entity name). Log that exact JSON array to the job output, then use it to determine which entities to analyze and their types. Reuse this exact JSON array for step 5 without mixing in non-JSON log lines.

4. For each selected entity:

   • Perform schema coverage analysis using the skill rubric.
   • Determine whether there are actionable testing gaps.

5. After completing analysis of all selected entities, persist all analysis timestamps in a single atomic write by running:

   go run ./scripts/schema-coverage-rotation record \
     --memory /tmp/gh-aw/repo-memory/schema-coverage-rotation/memory/schema-coverage-rotation/schema-coverage.json \
     --entities '<json>'
   

   where <json> is the full JSON array returned by step 3 (the select output), unchanged except for shell-safe quoting.

Issue creation rules

• Create one issue per analyzed entity only when actionable testing gaps exist.
• Never create more issues than ${{ needs.pre_activation.outputs.issue_slots_available }} for the current run.

Issue content must include:

• Entity name.
• Entity type (resource or data source).
• Entity implementation directory path.
• Attributes with no coverage.
• Attributes with poor coverage.
• Prioritized top 5 gaps (or fewer if less exist).
• Concrete acceptance-test additions that would close those gaps.

If an analyzed entity has no actionable gaps, do not create an issue for it.

If at least one entity was analyzed but none has actionable gaps, you MUST call noop with a short reason.

Issue title length guardrail

GitHub issue titles are limited to 256 characters total, including the title-prefix that create-issue prepends automatically.

• This workflow's prefix is "[schema-coverage] " (18 characters), leaving 238 characters for the title you provide.
• Before calling create-issue, verify that len("[schema-coverage] ") + len(your title) is ≤ 256.
• For this workflow, the agent-provided title should normally be only the Terraform entity name. Do not add extra prose such as "schema coverage gaps" to the title; that detail belongs in the issue body.
• If the natural title would exceed the limit, shorten the variable portion. Prefer the core identifier (pattern label, entity name, or base test name) over long descriptive phrases.
• Do not include markdown heading markers (#), emoji, or the prefix label redundantly in the title. The title field is plain text.

Dispatch

After creating all issues for this run (or if no issues were created), call the dispatch_code_factory safe output tool once with dispatch: true to dispatch the code-factory workflow for each created issue.