By gestrich
A Claude/Codex plugin for project planning: capture notes, navigate plans, run sprints, sync with Jira/Confluence/Slack, and review PRs against documentation.
Onboard an existing project to the project-planning plugin's conventions — survey what's already there (notes, plans, AGENTS.md, commit history hints of Jira/Confluence/Slack usage), then propose and apply a structured plan of changes (folders to create, files to move, AGENTS.md project-planning section to add, planning sources to declare, plugin skills to enable). Use this skill whenever the user says "bootstrap this project", "set up project planning here", "make this project work with the planning plugin", "onboard this repo to project planning", "get this project ready for the planning skills", "wire this up to project planning", or anything along the lines of "land me in a new repo and make it usable by these skills". Trigger even when the user does not name the skill explicitly — if they're asking how to get a fresh project into shape so the other planning skills (notes, plan, sprint, etc.) will work, this is the right home. Always shows the plan-of-changes for confirmation before writing anything; never silently edits.
Sweep the project's configured Slack channel for the last week, cross-reference against the project's documentation source (usually Confluence), and propose documentation updates — clarifications, recurring questions that deserve a permanent home, decisions that were made in Slack but never written down. Use this skill whenever the user says "update the docs", "what's missing from the docs", "review last week's Slack for doc updates", "what should we have written down", "sweep Slack for doc updates", "what doc updates do we need", "anything to land in Confluence from Slack", or otherwise asks for a recurring documentation hygiene pass against recent team discussion. Trigger even when the user does not name the skill — if they're asking what should be captured in the docs based on what's happened in chat, this is the right home. Read-only by default; offers to stage a doc edit (or a draft message to the owning team) only after explicit confirmation.
Capture raw voice transcripts and brain dumps into a project's notes/ folder verbatim. Use this skill whenever the user says things like "here's my brain dump", "voice transcript", "dump notes", "capture this", "drop this into notes", or pastes a visibly long, unstructured, first-person block of text that reads like a spoken transcript (rambling sentences, repeated thoughts, no headings, conversational filler). Trigger even when the user does not explicitly say "notes" — if the input looks like a raw transcript, this skill is the right home for it.
Show what the plan is for the current project — what's in motion, what's next, what's blocked — by reading whichever planning sources the project uses (Jira, Confluence, a local plans/ folder, Slack-as-context, or some combination). Use this skill whenever the user asks "what's the plan", "what's next", "where are we on X", "show me the plans", "what's the plan for <area>", "catch me up on this project", "what's the status", or anything that asks for a synthesized view of the project's direction. Also use this skill in reconciliation mode when the user asks "sync plans", "reconcile with Jira", "reconcile the docs", "what's drifted", "what should graduate", "is the local plan still aligned with Jira", or anything that asks to compare local plans/ against external sources of truth. Trigger even when the user does not name a specific source — figuring out which sources apply is the skill's job. The skill reads the project's AGENTS.md to discover which sources are configured and then loads only the matching reference documents in plugin/skills/plan/.
Review a GitHub pull request against the project's own documentation — both directions. Fetch the diff from the GitHub repo(s) named in the project's AGENTS.md, cross-reference the change against the documented conventions (Confluence pages, local docs, plan files declared in AGENTS.md), and produce two outputs in one pass: (a) a PR review surfacing where the diff diverges from documented conventions, and (b) doc-update suggestions for patterns or decisions in the diff that aren't reflected in the docs yet. Use this skill whenever the user asks "review this PR", "review PR #123", "is this PR aligned with our docs", "check this PR against the spec", "what's wrong with this PR", "does this PR match our conventions", "check the open PRs for doc drift", "sweep open PRs", or pastes a GitHub PR URL and asks for feedback. Trigger even when the user doesn't name the skill — if they're handing you a PR and asking what it means against the project's documented direction, this is the right home. Read-only by default; any PR comment or doc edit requires explicit confirmation. Local checkout is out of scope in v1 — diffs come from the GitHub API.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A plugin for Claude Code and OpenAI Codex that gives you a consistent gateway into any project for capturing notes, navigating plans, running sprints, syncing with external sources of truth (Jira / Confluence / Slack), and reviewing PRs against documentation. Drop the plugin into a new project, declare a small project-planning section in the project's AGENTS.md, and the skills figure out which sources apply and read only what's relevant.
Add the marketplace:
claude plugin marketplace add https://github.com/gestrich/ai-project-planning
Install the plugin:
claude plugin install project-planning@gestrich-project-planning --scope user
Use
--scope userto install system-wide for all projects.
Restart Claude Code if necessary.
Uninstall:
claude plugin uninstall project-planning@gestrich-project-planning --scope user
Add the marketplace:
codex plugin marketplace add gestrich/ai-project-planning
Open a Codex session and install via the /plugins menu — select project-planning → Install plugin.
There is no
codex plugin installCLI command yet; installation must be done through the interactive TUI.
After installing, the plugin can be toggled without the TUI by editing ~/.codex/config.toml:
[plugins."project-planning"]
enabled = true # set to false to disable without uninstalling
Uninstall: Remove the [plugins."project-planning"] entry from ~/.codex/config.toml and run:
codex plugin marketplace remove gestrich-project-planning
claude --plugin-dir ~/path/to/ai-project-planning/plugin
--plugin-dir points at the directory that holds .claude-plugin/plugin.json — that's ./plugin inside this repo, not the repo root.
Skills trigger from natural-language prompts. You don't need to name the skill — the descriptions are tuned to fire on the kinds of phrases you'd actually say.
| Skill | Example trigger | What it does |
|---|---|---|
notes | "here's my brain dump" / pasting a long transcript | Writes the transcript verbatim to notes/YYYY-MM-DD-<slug>.md. No cleanup, no rewriting. |
plan | "what's the plan?", "what's next?", "catch me up" | Reads AGENTS.md to find which sources apply (Jira, Confluence, local plans/, Slack-as-context), loads only matching reference docs, synthesizes a brief view. |
plan (reconciliation mode) | "sync plans", "what's drifted?", "what should graduate?" | Same skill, different mode — compares local plans/ against external sources and reports graduation candidates, drift, and stale items. |
sprint | "what should I do next?", "plan my sprint", "what's on deck?" | Acts as a scrum-master coach: gathers context, asks challenging questions, drafts a sprint file at plans/sprints/YYYY-MM-DD.md after confirmation. |
bootstrap | "bootstrap this project", "set up project planning here" | Surveys an existing project, proposes a plan-of-changes (folders, file moves, AGENTS.md section), applies it after confirmation. |
docs-update | "update the docs", "what's missing from the docs?" | Sweeps the configured Slack channel for the last week, cross-references the docs, proposes updates. Stages a write or drafts a message to the owning team. |
pr-review | "review this PR", "is this PR aligned with our docs?" | Fetches the PR diff, cross-references the project's documented conventions, produces both PR-side findings and doc-update suggestions. |
All skills are read-by-default. Any external write (Jira transition, Confluence edit, Slack send, PR comment, local file change) requires explicit confirmation.
A project opts into this plugin by adding a ## Project Planning section to its AGENTS.md. Free-form markdown — there's no enforced schema. The skills read it in natural language.
A minimal example:
## Project Planning
**Planning sources:** Jira (project key `EXAMPLE`, epic `EXAMPLE-100`), Confluence (space `EX`, root page id `123456`), and the local `plans/` folder for in-incubation ideas. Source of truth for shipped work is Jira/Confluence; `plans/` is supplementary.
**Slack channel:** `C0123456789` (canonical discussion channel).
**GitHub repos:**
- `acme-org/example-service` — primary backend
- `acme-org/example-web` — web client
**Plugin skills enabled:**
- `notes` — dump raw voice transcripts into `notes/`. Use when I paste a brain-dump.
- `plan` — show the current direction. Use when I ask "what's the plan" or "what's next".
- `sprint` — plan the week. Use when I ask "what should I do this week".
- `pr-review` — review PRs against the documented conventions. Use when I paste a PR URL.
npx claudepluginhub gestrich/ai-project-planning --plugin project-planningSkills for Python architectural best practices including Service Layer pattern, testing, and layer placement guidance.
Architectural patterns and best practices for Swift application development
Skills for Python architectural best practices including Service Layer pattern, testing, and layer placement guidance.
AI-driven XCUITest automation for interactive UI control and screenshot capture
Phased development workflow tools for Claude Code
Build and maintain an LLM-curated personal knowledge base in your project — Andrej Karpathy's LLM Wiki pattern, designed to scale to thousands of pages without becoming a context bottleneck. Now with an optional compiled graph layer for typed, provenance-backed relationships.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Connect to Atlassian products including Jira and Confluence. Search and create issues, access documentation, manage sprints, and integrate your development workflow with Atlassian's collaboration tools.
AI-powered wiki generator for code repositories. Generates comprehensive, Mermaid-rich documentation with dark-mode VitePress sites, onboarding guides, deep research, and source citations. Inspired by OpenDeepWiki and deepwiki-open.
20 modular skills for idiomatic Go — each under 225 lines, backed by 48 reference files, 8 automation scripts (all with --json, --limit, --force), and 4 asset templates. Covers error handling, naming, testing, concurrency, interfaces, generics, documentation, logging, performance, and more. Activates automatically with progressive disclosure and conditional cross-references.
Claude + Obsidian knowledge companion. Sets up a persistent, compounding wiki vault (Karpathy's LLM Wiki pattern). v1.7 "Compound Vault" + v1.8 methodology modes close 5 of 5 priority gaps from the May 2026 compass artifact. Ships: substrate alignment with kepano/obsidian-skills, default Obsidian CLI transport, hybrid retrieval (contextual prefix + BM25 + cosine rerank per Anthropic's Sept 2024 research), per-file advisory locking for multi-writer safety, pre-commit verifier agent, AND methodology modes (LYT / PARA / Zettelkasten / Generic) for first-class organizational support no other Claude+Obsidian competitor offers. v1.7.x audit closure: every BLOCKER + HIGH + MEDIUM + LOW finding from the v1.7.0 audit is CLOSED or DEFERRED-with-rationale. Optional DragonScale Memory extension (log folds, deterministic addresses, semantic tiling lint, boundary-first autoresearch).