Imported from diablodale/gpg-bridge (.agents/skills/codecov/SKILL.md). Install upstream with npx skills add diablodale/gpg-bridge --skill codecov. Copyright stays with the author.

Codecov Coverage Skill

Fetches and analyzes code coverage data from the Codecov API for a given repository.

Inputs Required

• owner: Codecov owner username (e.g. diablodale)
• repo: Repository name (e.g. gpg-bridge)
• task: What coverage information is needed (overall totals, file list ranked by coverage, missed lines in a specific file, PR comparison, etc.)
• branch or sha (optional): Scope to a specific branch or commit; defaults to head of default branch

Process

1. Load reference/api-endpoints.md to identify the correct endpoint and its query parameters for the task.

   • If deeper parameter/schema details are needed, consult reference/schema.yaml (OpenAPI 3.0 YAML, full spec)

2. Select the endpoint that best matches the task:

   • Overall coverage percentage → totals/
   • All files ranked by coverage → report/
   • Missed lines in one specific file → file_report/{path}/
   • PR or commit diff comparison → compare/

3. Make the API request using any available tool that supports HTTP REST with JSON. Refer to reference/api-endpoints.md for the base URL, required headers, and query parameter names.

4. Parse the JSON response. Extract only the fields relevant to the task.

   • For missed lines: filter line_coverage entries where status == 1 (miss; 0=hit, 1=miss, 2=partial).
   • For file rankings: sort by totals.coverage ascending.

5. Present findings concisely, prioritizing actionable information. Do not dump raw JSON unless explicitly requested.

Output

• Coverage totals: percentage, lines, hits, misses
• File list: table sorted by coverage ascending, showing file name, coverage %, and miss count
• Missed lines: list of line numbers with a note on what is uncovered
• PR comparison: coverage delta, impacted files, base vs head totals