agentic-contributor

Orchestrate an OSS-contribution task — find issues, research status, draft questions/feedback/new issues, learn norms, set up env (draft-only)

This agent checks whether a GitHub issue is free to claim. Use this agent when the orchestrator needs to: determine if an issue is already assigned or has an active linked PR, assess recent comment activity for informal claims, flag stale claims, and produce a clear "appears free to take" vs. "appears already claimed" verdict before the smart-questions skill drafts a claim comment. <example> Context: User found issue #4321 in apache/airflow and wants to claim it before starting work. user: Is apache/airflow issue #4321 free to take? assistant: I'll dispatch the oss-claim-analyst agent to check the assignee field, any linked open PRs, and recent comment activity on issue #4321, then give a clear verdict with the evidence. <commentary> The agent uses `gh issue view` and `gh pr list --search` to gather assignee, linked PRs, and recent comments. It reports the evidence concisely and either clears the issue or flags it as claimed, optionally noting stale activity. It never posts a claim comment itself. </commentary> </example> <example> Context: User wants to work on apache/spark issue #8765 but is worried someone already grabbed it. user: Check if spark issue #8765 is claimed before I spend time on it. assistant: I'll use the oss-claim-analyst agent to inspect assignees, linked open PRs, and the last 30 days of comment activity on issue #8765, then give a "free" or "claimed" verdict with supporting evidence. <commentary> The agent detects stale claims (claimed long ago, no recent activity) and flags them separately so the user can make an informed decision. It remains read-only throughout. </commentary> </example>

This agent fetches and analyzes GitHub data for open-source projects. Use this agent when the orchestrator needs to: retrieve project status (releases, issues, PRs), find candidate issues or PRs to contribute to, gather a PR's diff and CI/discussion context, look up contribution documents, or supply raw GitHub data to any other skill or command in this plugin. <example> Context: User wants the latest status of apache/airflow — recent releases, open issues, active PRs. user: What's the current state of apache/airflow? Any interesting open issues or recent releases? assistant: I'll use the oss-researcher agent to fetch the latest releases, recently-updated open issues, and recently opened/merged PRs from apache/airflow, then produce a timestamped status report with direct links. <commentary> The agent uses mcp__plugin_agentic-contributor_github__* read tools (or gh CLI fallback) to paginate through releases, issues, and pulls endpoints in small batches, then returns structured findings. </commentary> </example> <example> Context: User selected issue #12345 in apache/spark and wants to understand the linked PR before drafting feedback. user: Can you pull the diff, CI results, and discussion for apache/spark PR #9876? assistant: I'll dispatch the oss-researcher agent to fetch PR #9876's description, changed files, CI check status, and comment thread, then summarise where the PR stands and what kind of help would be most useful. <commentary> The agent prefers mcp__plugin_agentic-contributor_github__get_pull_request and related read tools, falls back to `gh pr view --json` if the MCP token is absent, and never posts or mutates. </commentary> </example>

This skill should be used when the user asks to learn how to contribute to an open-source project, understand a project's governance, find out about CLA or DCO requirements, learn the correct commit or PR title format, or find out where project decisions happen. Trigger phrases include: "how do I contribute to X", "what are the contribution norms", "do I need to sign a CLA", "what's the DCO", "where do decisions happen", "what's the governance model", "CONTRIBUTING guide", "how should I format my PR title", "what mailing list should I use".

This skill should be used when the user asks to set up their local development environment for an open-source project. Trigger phrases include: "set up the project locally", "how do I clone and build X", "help me get the dev environment running", "how do I run the tests for X", "I can't get the build to work", "set up apache/airflow locally", "set up apache/spark locally", "what are the prerequisites for X".

This skill should be used when the user asks to draft a bug report, file a new issue, report a bug in a project, draft a feature request, propose a feature, write up an issue for a project, or open an issue about a problem or improvement. Trigger phrases include: "draft a bug report", "file a new issue", "report a bug in X", "draft a feature request", "propose a feature", "write up an issue for X", "open an issue about". All output from this skill is a DRAFT for the user's review — never posted automatically.

This skill should be used when the user asks to find or match an issue or PR suited to them in a target open-source project. Trigger phrases include: "find an issue to work on", "match me to a task", "suggest a good first issue", "what can I contribute", "find a PR I can help with", "show me newcomer-friendly issues", and close variants. The skill runs after oss-researcher has gathered raw issue/PR data and applies the matching/ranking logic on top of it.

This skill should be used when the orchestrator needs to classify a user's open-source contribution intent into one of the nine supported scenarios. Trigger phrases include: "find an issue to contribute", "check project status", "draft a question to maintainers", "understand contribution norms", "set up the project locally", "claim an issue", "give feedback on a PR", "respond to PR review", "draft a bug report", "file a new issue", "propose a feature", and close variants of each.

1 hook across 1 event