Procoder plugin
Procoder is a single Go binary with a commit gate, quality controllers and a lessons ledger so agents cannot ship unfinished work
- 211 ★ on the repository
- Last commit
Extended access
Why this level
- Hooks fire on every write and session start, and `procoder init` installs the repository tools on the machine
- `procoder copilot-leak` can file GitHub issues after a terminal confirmation
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/procoderDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Hooks fire on every write and session start, and `procoder init` installs the repository tools on the machine
- `procoder copilot-leak` can file GitHub issues after a terminal confirmation
This is third-party code. Review the repository files before installing.
What it does
Procoder gives the agent senior developer checks. `procoder check` formats code in popular languages with one canonical formatter per language, finds merge conflict markers, junk and oversized files, secrets and AI-attribution lines, runs the linter and checks CI and documentation hygiene. A tool that did not run counts as failing: the `unchecked` status is said out loud. The quality chain does not let code in without thought. A spec closes design gaps and `spec check` blocks while sections are missing. A plan becomes tasks and `plan check` catches placeholders, while a backlog runs a project through milestones, epics and sprints. `todo close` refuses until criteria are checked and the gate is clean, and `procoder test` runs the real suite and tells not run from green. The engine computes and reports and the agent acts: Procoder never edits code on its own. Hooks fire on every write and session start. The lessons ledger turns every escaped bug into a linter rule or a test.
Who it is for. Developers who hand large tasks to an agent and want hard gates before commit, a plan and spec before code, since they do not trust the word done without a check.
Good fit when
- The agent says done while conflict markers, junk files and unformatted code reach the commit
- A large project runs by milestones, epics and sprints and closing tasks takes evidence
- A bug the reviewer missed becomes a linter rule or a test through the lessons ledger
- You want the same set of checks in Claude Code, Cursor and Codex
Not a fit when
- A one-off script or experiment: gates and specs only slow it down
- Someone else's repository without `.procoder/`: only secret, oversize and junk checks apply
- You want a tool that fixes files itself: Procoder only reports findings
Example request
Run procoder check on my changes and show the blocking findings before the commitLimitations
- The full check set applies in a repository that adopted Procoder: a `.procoder/` directory or an `AGENTS.md` naming it. In someone else's repository the short set runs, and every run states the mode.
- Test coverage is only reported and never enforced. Blocking on a red suite is switched on with `[test] policy = "block"`.
- The only command that changes files without the agent acting is `procoder prune --apply`: it removes superseded plugin copies from the cache and asks twice.
- `procoder copilot-leak` files GitHub Copilot auto-review findings as issues, but only after your yes in a terminal and after stripping traces of your code.
How to disable. For the Claude Code plugin run `/plugin uninstall procoder`. Hooks and adapters of other agents go with their files, and the project `.procoder/` directory is removed by hand.
FAQ
How do I install Procoder in Claude Code?
Run `/plugin marketplace add azrtydxb/procoder`, `/plugin install procoder` and `/procoder:init`. The last one installs the tools the repository needs.
Which agents are supported?
Claude Code, Cursor, Windsurf, Cline, Kilo Code, Roo, Kiro, Codex CLI, Copilot CLI, Gemini, OpenCode and any agent that reads `AGENTS.md`. Adapters are described on the project site.
Does Procoder edit code itself?
No. The binary computes findings and hands the agent the fixed content, and the agent applies it. The one exception is `procoder prune --apply`, which cleans the plugin cache.
What does unchecked mean?
A check tool did not run or gave no result. Procoder counts it as failing and says so, showing no green status.
How do self-serve and forced checks differ?
The agent runs `check`, `format`, `lint` and `scan` itself, while hooks fire on every write and session start and cannot be skipped.
Discussion
No questions or reviews yet.
Related
PluginsA 33-skill plugin with gates: question the request, verify facts against sources, test first, review by a second model
A plugin makes the agent explain its code and, before a push, ask questions so you understand what you are merging.
GigaCode project flows with hooks, gates and OpenSpec
gigacodeAgents
A monorepo of flows for the GigaCode CLI: development with stages and hard gates, a work progress observer and reverse analysis of code into a spec
A terminal coding agent in Go: GLM, DeepSeek and local Ollama models, 59 built-in tools and MCP
A Claude Code and Codex plugin with skills for code research, tests, refactoring and independent review of finished changes
A governance layer for agentic coding: enforced gates for tests, reviews, security and decisions, shared across Claude Code, Codex CLI and Pi
Comparison with similar tools
Cards that share the main search query, and how they differ.
README in short
The README presents Procoder as a binary with a gate the agent cannot talk past. It starts with a quick start and a `procoder check` example, then the agent-hook-binary diagram, the check inventory, the quality chain, the self-learning loop, the test domain and portability to every agent.