graph-wiki

Plugin graph-wiki from marketplace agent-research in psprowls/agent-research. Commands read from the repo's manifest — 1 month ago.

Skills (23)

• auto-drive

  Use when driving a work item's full pipeline unattended via Orca-supervised workers — an epic's entire dependency graph or a lone item's phase sequence — instead of walking one /graph-wiki:next stage at a time by hand. Runs a stateless plan/act/wait coordinator loop over `gw work orchestrate` and th

• brainstorming

  You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

• checking-gates

  Use when picking up a user-gate task OR when a hook demands re-validation. Runs the "do I know HOW?" self-check; if the HOW is clear, executes the verification and posts evidence; if not, hands off to specifying-gates. Kept deliberately separate from executing-plans so that without the opt-in hook,

• dispatching-parallel-agents

  Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

• executing-plans

  Use when you have a written implementation plan to execute in a separate session with review checkpoints

• finishing-a-development-branch

  Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup

• finishing-relay

  Use when the finish stage of a work item is dispatched under auto-drive with mode relay — sends one orca orchestration ask carrying the merge/PR/hold/discard decision instead of finishing-a-development-branch's interactive menu, executes the chosen outcome, and settles the item. Dispatched by the gr

• graph-wiki

  Use when building or maintaining a persistent wiki alongside any source-code project — single packages, monorepos, or hybrid shapes. Builds a code graph and renders one page per entity (repository, package, app, agent_plugin, dependency, test_suite) into a single entities/ folder. Triggers include "

• planning-epics

  Use when dispatched as the `plan` stage of an epic work item — decomposes the epic into concrete child work items, runs dependency/sequence analysis, files each child via `gw work file --parent/--depends-on`, and writes the plan_doc. Sibling of writing-plans (which plans a single feature).

• receiving-code-review

  Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation

• requesting-code-review

  Use when completing tasks, implementing major features, or before merging to verify work meets requirements

• specifying-gates

  Use when a user-gate task has requiresUserSpecification=true OR the agent's "do I know HOW?" self-check returns no. Locks down verification mechanics through a short AskUserQuestion sequence and writes the answers back into the task's metadata. Does NOT run the verification itself.

• subagent-driven-development

  Use when executing implementation plans with independent tasks in the current session

• systematic-debugging

  Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

• test-driven-development

  Use when implementing any feature or bugfix, before writing implementation code

• using-git-worktrees

  Use when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback

• using-workflow

  Use when starting any conversation - establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions

• verification-before-completion

  Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

• workflow

  Use when driving a work item through its development pipeline — runs `gw work next` to compute the stage, dispatches the stage skill (brainstorming, systematic-debugging, writing-plans, subagent-driven-development, test-driven-development, finishing-a-development-branch), verifies the artifact, and

• writing-clearly-and-concisely

  Apply Strunk's timeless writing rules to ANY prose humans will read—documentation, commit messages, error messages, explanations, reports, or UI text. Makes your writing clearer, stronger, and more professional.

• documentation

  Creates, structures, and reviews technical documentation following the Diátaxis framework (tutorials, how-to guides, reference, and explanation pages). Use when a user needs to write or reorganize docs, structure a tutorial vs. a how-to guide, build reference docs or API documentation, create explan

• writing-plans

  Use when you have a spec or requirements for a multi-step task, before touching code

• writing-skills

  Use when creating new skills, editing existing skills, or verifying skills work before deployment

Slash commands (16)

• /graph-wiki:archive

  Archive terminal-status work items (resolved/wontfix/superseded) — sweep mode by default, or target specific slugs. Presents the plan and asks for confirmation before executing. Invokes `gw work archive`. Usage /graph-wiki:archive [slug...]

• /graph-wiki:auto-drive

  Drive a work item's full pipeline unattended via Orca-supervised workers — an epic's DAG or a lone item's phase sequence — dispatching each ready stage as a fresh Orca worker, handling attend (design) and relay (finish) human gates, and resuming cleanly after a restart. Re-running the same command i

• /graph-wiki:bootstrap

  Bootstrap a fresh Code Wiki in the resolved graph-wiki workspace — schema files and starter templates. Wiki is created at <workspace>/wiki/. Usage /graph-wiki:bootstrap --topic "<topic>" [--tool all|claude-code|codex|cursor|antigravity]

• /graph-wiki:file

  Interactively file a new work item into the wiki — gathers title, kind, summary, and affects conversationally, then invokes `gw work file` with the assembled values. Usage /graph-wiki:file

• /graph-wiki:gate-check

  Run the 'do I know HOW?' self-check on a user-gate task and either execute its verification with captured evidence, or hand off to /specify-gate when the HOW is ambiguous. Dormant unless the user-gate opt-in hook is registered.

• /graph-wiki:ingest

  Ingest a source file from raw/ into the Code Wiki — read, discuss, write summary, link relevant code entities via [[entities/...]] and update concept/ADR pages, propose ADRs if decisions are captured, flag contradictions with code, update index, append to log. Usage /graph-wiki:ingest <path-to-sourc

• /graph-wiki:lint

  Run a health check on the Code Wiki — mechanical (orphans, broken links, stale pages, missing frontmatter, duplicates, log gap), semantic (contradictions, cross-reference gaps, stale claims, ADR chain), code-drift (packages on disk vs. in vault, exports mismatch), and work lifecycle (31 rules for wo

• /graph-wiki:log

  Show recent entries from the Code Wiki log (<workspace>/wiki/log.md). Uses the standardized ## [YYYY-MM-DD] header format so grep + tail works. Usage /graph-wiki:log [--last N] [--op scan|ingest|query|lint|...]

• /graph-wiki:next

  Drive a work item through its pipeline — compute the next stage with `gw work next`, dispatch the stage skill, verify the artifact, advance with `gw work advance`. One stage per invocation; clear context and re-run between stages. Usage /graph-wiki:next <slug> [--descend]

• /graph-wiki:onboard

  Guided setup for graph-wiki workflow's optional features. Asks short multiple-choice questions and applies each answer via the gw config CLI. Terminal equivalent: gw config init. Manual equivalent: see README.md.

• /graph-wiki:proposals

  Review and dispose of curated-page proposals in `wiki/proposals/` — accept / reject / supersede. Approving only flips status; this skill then fans out one subagent per accepted proposal to author the destination page, flips notes to `created`, regenerates indexes, and archives. Usage /graph-wiki:pro

• /graph-wiki:query

  Query the Code Wiki — reads index.md first, drills into 3-10 relevant pages (concepts, entities, ADRs, sources), synthesizes answer with inline [[wikilinks]] and `code-paths:line`, and offers to file the answer back as a concept page. Usage /graph-wiki:query "<question>

• /graph-wiki:regen-index

  Rebuild wiki/work-index.json from the current wiki/work/*.md state. Run after filing or archiving work items, or when `gw work status` reports a missing sidecar. Invokes `gw work regen-index`. Usage /graph-wiki:regen-index

• /graph-wiki:scan

  Build the code graph and write one page per admitted entity (repository, package, app, agent_plugin, dependency, test_suite) into the wiki's single entities/ folder. Reports created/updated/deleted entities by URI; surfaces deletions for confirmation. Workspace and repo discovered automatically. Usa

• /graph-wiki:specify-gate

  Lock down verification mechanics for a user-gate task whose HOW is ambiguous. Asks 4 short questions, writes the answers back into the task's metadata, then returns control to executing-plans. Dormant unless the user-gate hook is registered.

• /graph-wiki:status

  Show a one-screen work item rollup — counts by status/kind, in-flight items, stuck items. Hints to run regen-index when the sidecar is missing or stale. Invokes `gw work status`. Usage /graph-wiki:status

Subagents (4)

• ingestor

  Dispatched sub-agent that ingests a source file from raw/ into the Code Wiki. Reads the source, proposes TL;DR and key claims, identifies which code entities and concepts will be touched, flags contradictions with wiki or code, proposes ADRs when decisions are captured, and — after user confirmation

• librarian

  Dispatched sub-agent that answers queries against a Code Wiki. Reads index.md first, drills into 3-10 relevant pages across categories (concepts, entities, ADRs, sources, work), synthesizes an answer with inline [[wikilink]] and `code-path:line` citations, and offers to file the answer back as a new

• linter

  Dispatched sub-agent that runs a health check on a Code Wiki. Mechanical checks via scripts (orphans, broken links, stale pages, missing frontmatter, duplicate titles, log gaps, CODE DRIFT), semantic checks (contradictions vault↔vault and vault↔code, stale claims, concept gaps, ADR chain health, cro

• scanner

  Dispatched sub-agent that walks the monorepo, builds the code graph, and writes one graph-derived page per admitted entity into the wiki's single `entities/` folder (repository, package, app, agent_plugin, dependency, test_suite). Reports added/updated/deleted entities by URI and surfaces deletions

Hooks (5)

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}/hooks/run-hook.cmd" session-start

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}/hooks/run-hook.cmd" skill-doc-routing

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}/hooks/run-hook.cmd" pre-taskcreate-model-tier

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}/hooks/run-hook.cmd" pre-agent-model-routing

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}/hooks/run-hook.cmd" pre-askuser-handoff-guard

Similar plugins