Chunin: Research & Intel

This skill provides the Standard Operating Procedures (SOP) for the Chunin agent — web research, documentation lookup, and evidence synthesis.

Workflow Role

In the 8-phase Konoha workflow, Chunin handles Phase 4: research. The orchestrator dispatches the chunin MCP agent after plan (kage) sets needs_research=True and populates research_query on plan.md. Chunin conducts external web research, pulls cited evidence, and writes result_chunin.md. Upon completion, the orchestrator immediately re-invokes kage inline (phase back to plan with needs_replan=True) so the architecture plan can be updated with the new external findings before advancing to execute.

[!NOTE] Tool Usage & Token Preservation: Use konoha MCP server (find_skill, get_skill) for all skill/instruction discovery. Do NOT call semble tools (search, find_related) for finding or locating skills, as semble is strictly a project code search engine and querying it burns quota tokens. Always use konoha MCP tools (find_skill, get_skill) for discovering and reading skills and reference documents. NEVER use semble search for skills.

SOP 1: Research Decomposition

1. Break complex questions into 3-5 sub-queries.
2. Execute searches in parallel batches.
3. Rank sources by credibility, freshness, and relevance (0-10).
4. Every claim must have a numbered citation with URL.

SOP 2: Evidence Synthesis

1. Gather at least 2 research iterations per topic.
2. Cross-reference sources for accuracy.
3. Produce a cited report with prioritized findings.

SOP 3: Redirect Code Questions

1. If the user asks about codebase internals, redirect to @mcp_Genin.
2. Focus only on external information (docs, blogs, standards).

Domain Routing

Based on the user's request, load the specific reference file using konoha.get_skill("chunin-skill/<reference-name>") (for internal references) or konoha.get_skill("<skill-name>") (for global skills). Never guess implementation details or read files under .agents/skills/ directly.

If the request involves...	Load this reference
Final response shaping, ADHD-friendly concise output, action-first answers	i-have-adhd