MCP server that gives Claude Code the rationale behind a symbol before editing - synthesized from git history, PRs, and call-graph context.
This repository also configures its own agents. See what whygraph tells them →
Skill Claude Code
Fetches WhyGraph rationale - intent, constraints, tradeoffs, and risks - for existing code before it is changed, so an edit respects the original intent instead of rediscovering it.
Skill Claude Code
Show the WhyGraph rationale for a code chunk - purpose, why, constraints, tradeoffs, risks. Read-only lookup.
Skill Claude Code
Execute a reviewed WhyGraph plan (or a split's tasks.json) one step per fresh implementor at the step's model, with a cached-card review gate and a spec evaluation after each step, a separate acceptance run at the end, then quality and UI evaluations. The plan tool does the bookkeeping, a counted recovery ladder…
Skill Claude Code
Plan a non-trivial code change with WhyGraph rationale + CodeGraph symbols. Runs a fan-out/fan-in pipeline (planner -> one researcher per knowledge source -> synthesizer), splits a large task into child plans only when needed, and asks the few plan-shaping questions after research.
Skill Claude Code
Check a diff against the recorded WhyGraph rationale - per hunk, whether a constraint or risk on the enclosing symbol's card is at stake. Reads cached cards only, no LLM spend, unless --generate is passed.
Skill Claude Code
Propose a WhyGraph project note - a decision, constraint, gotcha or convention anchored to code - for a person to confirm in the portal.
Skill Claude Code
Capture screenshots, accessibility snapshots and console and network signals of the app's routes with the bundled capture script, then print the contact sheet path. Uses the UI profile that /whygraph-plan writes.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: