mpt-ext-task-run-repository-checks

name	mpt-ext-task-run-repository-checks
description	Run the repository-required local validation flow to verify changes are ready for commit or review. Determines the right checks for the changed scope, runs them in order, and reports failures.

Run Repository Checks

Purpose

Run the repository-required local checks and tests for the current change scope before commit or review handoff.

Use When

• The user wants to verify that current changes pass the required local validation flow.
• The task requires running repository checks before commit.
• The task requires rerunning validation after review-driven fixes.
• The repository defines local checks through repo docs, Makefile targets, or shared validation guidance.

Do Not Use When

• The task is only to create a branch or update Jira status.
• The task is only to create a commit without running or deciding checks.
• The task is only to open or update a pull request.
• The task is to invent a validation flow without reading repository docs first.

Inputs

• Current repository change scope.
• Repository validation rules from repo docs.
• Optional explicit scope from the user when only part of the repository should be validated.
• Installed shared package root:

${MPT_EXTENSION_SKILLS_HOME:-$HOME/.mpt-extension-skills}/current

Assumptions

• The target repository is available locally and the documented validation commands can be executed in the current environment.
• Repository-required tooling and dependencies are installed or otherwise available before the check run starts.
• Any credentials, tokens, or repository access needed by the documented validation flow are already available.

Shared References

Use this shared document as the source of truth instead of restating its guidance. When shared guidance is needed, resolve it from ${MPT_EXTENSION_SKILLS_HOME:-$HOME/.mpt-extension-skills}/current when available; otherwise read the same path from the main branch of the shared GitHub repository.

• knowledge/build-and-checks.md

Workflow

1. Build repository context first.

• Read the target repository AGENTS.md once per session. If you already loaded it earlier in this session and still have its full contents, reuse them instead of re-reading; if the context was summarized or you are unsure it is complete, read it again. Do not pre-load shared docs in this step; read them lazily only when the repository points to them.
• Read repository-specific docs when they exist, because they may extend or override shared guidance.
• Read shared docs only when the repository explicitly points to them. Resolve those shared docs from ${MPT_EXTENSION_SKILLS_HOME:-$HOME/.mpt-extension-skills}/current when available; otherwise read them from the main branch of the shared GitHub repository.

2. Resolve the required validation flow.

• Determine which local commands are required for the current change scope.
• Use repository docs as the source of truth for command names and ordering.
• For repositories that follow this shared package validation guidance, read knowledge/build-and-checks.md using the shared-doc resolution rule from the repository context step, and use it with related repository docs as the shared reference.

3. Prepare the environment only as required by the repo workflow.

• Confirm prerequisite tooling required by the repository validation flow.
• If the repository requires environment refresh after dependency-lock changes, include that step before running checks.
• If the repository relies on shared validation guidance from this package, read knowledge/build-and-checks.md using the shared-doc resolution rule from the repository context step, and use it as the source of truth for commit-oriented preparation such as pre-commit.
• Do not run unrelated setup steps that are not required for the current validation flow.

4. Run the required checks in order.

• Execute the repository-required validation commands in the documented sequence.
• Keep command output attributable to the specific check that produced it.
• If the user asked for a scoped validation run and the repository supports that safely, limit execution to that scope.

5. Interpret failures clearly.

• Separate environment or setup failures from real lint, test, or build failures.
• Treat a validation tool failing on its own configuration (for example an "unknown rule" or "unrecognized selector" error instead of a finding tied to a specific file or line) as an environment/setup failure, not a real finding; see the stale-toolchain pattern in knowledge/build-and-checks.md.
• Identify which command failed and which files or checks are implicated.
• If checks modify files automatically, report that clearly so the next task can decide whether to keep and recommit the changes.

6. Report the validation result.

• State which commands were run.
• State whether the validation passed fully, failed, or was only partially run.
• Show the blocking failures clearly when validation does not pass.

Guardrails

• Never guess the validation commands without reading repository context first.
• Never silently skip a required check in the repository-defined flow.
• Never present a partial validation run as if it were the full repository check set.
• Never hide auto-fixed file changes caused by validation tools.
• Never mix commit creation, PR creation, or Jira transitions into this task.

Expected Outcome

The repository-required local validation flow for the current change scope is run in the documented order, with a clear pass/fail result and precise reporting of any blockers or auto-generated file changes.