RenderCV Issue Solver
Picks up a GitHub issue on the rendercv repo, learns the codebase, ships a test-first fix, and opens a pull request.
The skill's instructions are non-malicious; permissions are proportionate to the task (gh issue reads, git branch, push, PR creation), with no credential theft or covert exfiltration; MIT license and traceable origin rendercv/rendercv. Deducted for: it autonomously picks an issue, creates a branch, pushes to origin, and opens a PR without any user-confirmation step — automated write operations with external effects lacking confirmation, isolation, or rollback guidance, and without stating assumptions about gh/git credentials and push targets.
Instructions are self-consistent with clear steps (red test first, all just format/check/test gates before proceeding) and basic failure-path guidance. Deducted for: the referenced development/testing context skills are not present in evidence so key paths cannot be statically confirmed; no failure feedback is defined for missing gh auth, no suitable issue, or broken environment.
The target scenario is clear (resolve a GitHub issue and open a PR for the rendercv repo), with explicit trigger logic. Deducted for: narrow applicability limited to rendercv contributors with no declared non-fit boundary; hidden dependencies on gh CLI, uv, just, and reachability of github.com — a risk for mainland-China users that is not disclosed.
Well-structured, progressive disclosure, complete command examples, consistent naming (claude/issue-<n>); the repo has MIT license, version 2.8, changelog link, and a named maintainer. Deducted for: the skill itself has no version/update record; its two dependency skills are unverified; README notes skills are auto-generated, implying doc-drift maintenance risk.
The workflow design is sound (failing test for bugs, full verification gates, structured PR body) and should yield a directly reviewable PR, giving clear marginal value. Deducted for: static review cannot verify output quality; success depends entirely on the absent context skills and the repo's test suite, so completeness and comparative benefit remain unevidenced.
The repo offers auditable material (CI badge, coverage badge, pytest/coverage config, promptfoo eval files), but these target the main rendercv code or a different skill, not this skill's key paths; no execution evidence exists in a static read, so only limited credit is given with no upward adjustment.
- The skill autonomously picks an issue, pushes a branch, and opens a PR to rendercv/rendercv without user confirmation — an external write operation; verify credential scope and prepare rollback (closing the PR) before use.
- The skill depends on two context skills (development/testing context) not present in the evidence; their existence and quality are unverified.
- The workflow relies entirely on github.com and gh CLI reachability, which may be unavailable from mainland-China networks.
- This is a static source review; no commands were executed, so all reliability conclusions are inference rather than measurement.
- The publisher is not verified by the FollowSkills curated registry and is treated as unknown.
What does this skill do, and when should you use it?
This skill turns 'solve an open-source issue' into a repeatable eight-step workflow: select an issue, understand the code, branch, reproduce the bug, implement the fix, verify, commit and push, then open a PR. It mandates a failing test before any bug fix and requires every verification command to pass, with coverage held at 100%. It is purpose-built for the rendercv/rendercv repository and relies on two sibling skill files in the same repo for development and testing context.
Reads GitHub issues via gh issue list and gh issue view, auto-selecting high-priority ones (skipping wontfix/question labels, preferring bugs and smaller scopes); creates a claude/issue-<number> branch from origin/main; for bugs, writes a failing test first and confirms it goes red with uv run --frozen --all-extras pytest; implements the fix following the rendercv-development-context and rendercv-testing-context skills in the same repo; runs just format, just check, just test, and just test-coverage as full verification; stages only changed files, commits, and pushes; finally opens a PR to main with gh pr create and reports the root cause, approach, tests, coverage status, and PR link.
- Maintainers who want an AI agent to work through rendercv bug reports to repo standards, producing review-ready PRs
- Contributors with a specific issue number who want the full test-first workflow plus a well-structured PR description
- Teams enforcing a rule that every fix ships with a failing test and 100% coverage, using this skill to apply the discipline
- Newcomers to the rendercv project using the workflow to learn the codebase structure and verification commands
What are this skill's strengths and limitations?
- Complete, rigorous workflow: explicit commands for every step from issue selection to PR creation
- Enforces test-first development: no bug fix proceeds without a red failing test
- High quality bar: format, check, test, and coverage must all pass, with coverage held at 100%
- Well-structured PR descriptions with summary, change list, and test plan
- Works only against rendercv/rendercv; not directly reusable for other projects
- Depends on two sibling skill files in the same repo for context, so it cannot run when copied in isolation
- Requires gh, git, uv, and just preinstalled; the skill does not explain how to set these up
- Pushes branches and opens PRs autonomously, so you must trust its automatic issue selection
How do you install this skill?
The skill is part of the skill collection in the rendercv/rendercv repo (the repo README gives npx skills add rendercv/rendercv-skill as the collection install command). The skill file lives at .claude/skills/solve-rendercv-issue/SKILL.md; the source does not document standalone installation steps for this skill alone. The rendercv-development-context and rendercv-testing-context skills in the same repo are prerequisites.
How do you use this skill?
In an environment with gh, git, uv, and just, prompt your agent with an issue number or URL, e.g.: 'Use the solve-rendercv-issue skill to resolve issue #123 in the rendercv repo.' With no number given, the skill lists open issues and picks the most impactful one automatically. When done, it reports the root cause, the fix approach, test changes, coverage status, and the PR link.
How does this skill compare with similar options?
The same repo bundles five sibling skills, including rendercv-development-context and rendercv-testing-context, which supply development and testing context; this skill is the end-to-end execution flow that references rather than replaces them. The source names no external alternatives.