Identify the current branch and its associated PR. If no PR exists, work directly from the branch name.
Fetch CI/CD run results using gh run list and gh run view to find failing jobs on the current branch. If a PR exists, also check gh pr checks for a consolidated view.
For each failing job, fetch the full logs with gh run view --log-failed to extract the root cause. Categorize failures by type:
- Lint / formatting errors: fixable automatically.
- Type errors: fixable automatically.
- Test failures: investigate whether the failure is in the test itself or in the code under test. Fixing a flawed test assertion is higher risk — flag it for human approval before modifying tests.
- Build errors: investigate missing dependencies, config issues, or breaking changes.
- Infrastructure / environment errors (flaky network, missing secrets, etc.): flag for human — do not attempt to fix.
Present a summary to the human: what failed, why, and your proposed fix for each. Explicitly wait for human approval before proceeding. Do not spin up agents until the human has confirmed the plan.
Spin up an agent or team of agents (as appropriate for scope) to implement the approved fixes. Give each agent a focused task with a clear success criterion. No need to create a new branch or worktree — fixes go on the current branch.
Once fixes are implemented, run the full project checks locally to verify. Note: per-file lint and type checks run automatically via hooks on every edit — focus this step on full-project commands (e.g. full typecheck, test suite) that hooks defer. Discover the correct commands from AGENTS.md, package.json, or pyproject.toml in the affected repo — do not guess.
Run /critique to review, commit, and push the fixes.
Monitor the new CI run with gh run watch until it completes. If new failures appear, repeat from step 3.

Más de este repositorio

mismo repositorio

comment-linear

Luney-Industries/farty-bobo

Post or update a comment on a Linear ticket using the Linear MCP server. Accepts a ticket ID (e.g. ENG-123), a Linear ticket URL, or a direct link to an existing comment. Works for both human-authored comments and agent-generated summaries.

2026-07-174

create-linear-document

Luney-Industries/farty-bobo

Creates a Linear document (the Documents feature, distinct from issues/projects) from context in the current conversation, a plan, or a file. Use when the user wants to write up a doc, spec, RFC, or notes page in Linear and attach it to a team, project, initiative, cycle, or issue.

2026-07-164

critique

Luney-Industries/farty-bobo

Used to run githooks, perform code & plan reviews with expert critics, and commit/push verified changes

2026-07-104

create-linear-project

Luney-Industries/farty-bobo

Creates a Linear project under an initiative from a draft list of tickets, clarifies ambiguity with the human, then creates the project and each ticket via /create-linear-ticket. Use when the user wants to stand up a new project in Linear from a plan, doc, or ticket list.

2026-07-104

build

Luney-Industries/farty-bobo

Used to build and execute an approved implementation plan by spinning up agents

2026-06-304

plan-task

Luney-Industries/farty-bobo

Used to gather requirements, clarify scope, and create implementation plans for new feature requests or bug fixes

2026-06-304

Investigates and resolves CI/CD failures for the current branch or PR

Identify the current branch and its associated PR. If no PR exists, work directly from the branch name.

Fetch CI/CD run results using gh run list and gh run view to find failing jobs on the current branch. If a PR exists, also check gh pr checks for a consolidated view.

For each failing job, fetch the full logs with gh run view --log-failed to extract the root cause. Categorize failures by type:

Lint / formatting errors: fixable automatically.
Type errors: fixable automatically.
Test failures: investigate whether the failure is in the test itself or in the code under test. Fixing a flawed test assertion is higher risk — flag it for human approval before modifying tests.
Build errors: investigate missing dependencies, config issues, or breaking changes.
Infrastructure / environment errors (flaky network, missing secrets, etc.): flag for human — do not attempt to fix.

Present a summary to the human: what failed, why, and your proposed fix for each. Explicitly wait for human approval before proceeding. Do not spin up agents until the human has confirmed the plan.

Spin up an agent or team of agents (as appropriate for scope) to implement the approved fixes. Give each agent a focused task with a clear success criterion. No need to create a new branch or worktree — fixes go on the current branch.

Once fixes are implemented, run the full project checks locally to verify. Note: per-file lint and type checks run automatically via hooks on every edit — focus this step on full-project commands (e.g. full typecheck, test suite) that hooks defer. Discover the correct commands from AGENTS.md, package.json, or pyproject.toml in the affected repo — do not guess.

Run /critique to review, commit, and push the fixes.

Monitor the new CI run with gh run watch until it completes. If new failures appear, repeat from step 3.