oss-scenario-routing

name

oss-scenario-routing

description

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.

OSS Scenario Routing

Classify the user's intent from $ARGUMENTS (or from the clarifying answer) into exactly one of the nine scenarios below, then route to the correct skill and subagent.

Scenario Catalog

#	Scenario key	Intent signals	Skill to load	Agent to dispatch	Output type
1	status	"what's new", "latest releases", "recent issues/PRs", "project news", "what's happening in X"	—	oss-researcher	Timestamped status report saved to .oss-drafts/
2	find	"find an issue", "match me to a task", "good first issue", "what can I work on", "suggest a PR to help with"	issue-matching	oss-researcher	Ranked shortlist with rationale saved to .oss-drafts/
3	norms	"contribution norms", "how do I contribute", "CLA", "DCO", "CONTRIBUTING", "governance", "how decisions are made"	contribution-norms	oss-researcher	Contribution briefing saved to .oss-drafts/
4	setup	"set up locally", "clone and build", "dev environment", "run the tests", "how do I build X"	dev-env-setup	oss-researcher (optional, read-only — to fetch CONTRIBUTING/setup docs)	Step-by-step guided walkthrough saved to .oss-drafts/
5	clarify	"draft a question", "ask the maintainer", "clarify the issue", "what does this issue mean", "I don't understand the design"	smart-questions	oss-researcher	DRAFT questions saved to .oss-drafts/ (review-before-send)
6	claim	"claim an issue", "can I take this", "is this issue free", "I want to work on #N"	smart-questions	oss-claim-analyst	Claim assessment + DRAFT claim comment saved to .oss-drafts/
7	engage	"give feedback on a PR", "help with someone's PR", "comment on PR", "review an in-progress PR", "there's a linked PR"	smart-questions	oss-researcher	DRAFT feedback comment saved to .oss-drafts/
8	review-reply	"respond to review", "reply to review comments", "maintainer left feedback on my PR", "how do I respond to this review"	smart-questions	oss-researcher	DRAFT review replies saved to .oss-drafts/
9	report	"report a bug", "file an issue", "draft a bug report", "draft a feature request", "propose a feature", "write up a new issue"	issue-drafting	oss-researcher	DRAFT new issue (bug/feature) saved to .oss-drafts/

Classification Rules

To classify the user's intent:

1. Match the $ARGUMENTS text against the intent signals column (exact phrases, close synonyms, or paraphrases). The first strong match wins.
2. If no strong match is found, ask one clarifying question: present the scenario menu (numbered list of the 9 scenarios) and ask the user which applies.
3. Never ask more than one clarifying question. After the answer, classify and proceed.

Claim → Engage Branch

After dispatching oss-claim-analyst for the claim scenario:

• If the analyst's verdict is "appears free to take": proceed with the smart-questions skill to draft a polite claim comment.
• If the analyst's verdict is "appears already claimed" (assignee, active PR, or recent intent comment): automatically switch to the engage scenario — load smart-questions and dispatch oss-researcher to gather the linked PR context and draft constructive feedback instead.

Routing Steps

Once the scenario is classified, the /oss command must:

1. Load the mapped skill (by naming it) to apply domain-specific guidance.
2. Dispatch the mapped subagent via the Task tool with:
   • The target project (owner/repo).
   • The specific item (issue number, PR number, or query intent).
   • The scenario key so the subagent knows its focus.
3. Produce the scenario's output as specified in the output-type column.

DRAFT-ONLY Rule

All 9 scenarios save their output to .oss-drafts/ in the user's working directory. The orchestrator uses the Write tool to persist the file; subagents remain read-only.

Report tier (scenarios 1–4)

The orchestrator saves the output file and shows:

Saved to <path>. Review and edit this report file as needed before using it.

These scenarios never produce outbound text and never trigger any write action toward GitHub.

Draft tier (scenarios 5–9)

The orchestrator saves the draft file and shows the complete draft in chat, then adds:

DRAFT — saved to <path>. Review and edit this file before sending. This plugin will NOT post, comment, push, or send anything. Sending is handled by the separate execution/submission plugin.

• Always present the drafted content (questions, claim comment, feedback, review replies) for the user's explicit review.
• NEVER call any mutating tool, MCP write method, gh posting command, or git push. The PreToolUse guardrail hook will block such attempts regardless.
• Writing local .oss-drafts/ files with the Write tool does NOT violate draft-only — local file writes cannot reach GitHub.