Skillget

SaneProcess

skill
Guvenlik Denetimi
Uyari
Health Uyari
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 7 GitHub stars
Code Gecti
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Workflow enforcement for coding agents: Claude Code hooks, Codex AGENTS/skills/MCP, and shared guardrails for safer, repeatable AI-assisted development.

README.md

SaneProcess

SaneProcess

Stop coding agents from wasting your time and crashing your machine.

SaneProcess is workflow enforcement for Claude Code, Codex, and compatible coding agents. It combines Claude-native hooks with Codex-native instructions, skills, MCP, and shared shell/script guardrails so the same SOP survives across clients.

452 tests. MIT licensed. Ruby. macOS + Linux.

Documentation Layout

SaneProcess is the source of truth for the anti-fragmentation rule used across SaneApps.

  • Update an existing canonical doc before creating a new one.
  • The normal app-repo skeleton is README.md, ARCHITECTURE.md, DEVELOPMENT.md, PRIVACY.md, and SECURITY.md.
  • If a repo needs extra docs, the owning canonical doc and the README must point to them clearly.
  • Public site pages should live in one obvious folder and the README should say where that folder is.

See DEVELOPMENT.md for the operating procedure behind this rule.

Client Support

SaneProcess is intentionally split into stable layers instead of pretending every client exposes the same API.

Client Native surface Stable SaneProcess path
Claude Code Lifecycle hooks (SessionStart, UserPromptSubmit, PreToolUse, PostToolUse, Stop, plus newer subagent events) scripts/hooks/*.rb + .claude/settings.json + MCP + shared guards
Codex AGENTS.md, .codex/config.toml, .agents/skills, MCP, approvals/sandbox/config AGENTS.md + .agents/skills + MCP + shared guards + SaneMaster.rb
Other agents Varies by tool AGENTS.md + repo scripts + git/pre-commit checks + shared shell/script guards

Important: Codex now documents an experimental features.codex_hooks flag, but it is still under development and off by default. SaneProcess does not rely on that as the primary path today.

The Problems

1. Orphaned processes eating your RAM

Claude Code and Codex both spawn agent and MCP sidecars that can outlive their parent sessions. They accumulate silently until your machine crawls or crashes.

SaneProcess kills them automatically on every session start — without touching your active sessions.

🧹 Cleaned up 3 orphaned Claude sessions
🧹 Cleaned up 7 orphaned MCP daemons
🧹 Cleaned up 2 orphaned subagents

It uses process tree traversal (BFS) to identify which processes belong to your current session and leaves them alone. Only orphans whose parent sessions have died get cleaned up.

2. Doom loops burning tokens

Claude guesses the same broken fix over and over. You watch it burn through tokens on the same error 10 times.

SaneProcess trips a circuit breaker after 3 consecutive failures or 3 identical error signatures. All edit operations are blocked until you acknowledge the problem.

🔴 CIRCUIT BREAKER TRIPPED
   3 consecutive failures with same error signature
   Say "reset breaker" after fixing the root cause.

The breaker persists across session restarts — Claude can't bypass it by restarting.

3. Edits without research

Claude assumes APIs exist without checking. It writes code using methods that don't exist, then fails, then tries a different nonexistent method.

SaneProcess blocks all edits until research is done across the required categories:

🔴 BLOCKED: Research incomplete
   ✅ docs   ✅ web   ❌ github   ❌ local
   Complete all required categories before editing.

Read-only tools (Read, Grep, Glob, search) are never blocked. The gate only applies to mutations.

Install

# Clone the repo
git clone https://github.com/sane-apps/SaneProcess.git

# Run the installer from your project directory
cd /path/to/your-project
/path/to/SaneProcess/scripts/init.sh

The installer copies hooks into your project's scripts/hooks/ directory, creates .claude/settings.json with Claude hook registration, seeds a starter AGENTS.md when missing, and mirrors repo skills into .agents/skills/ for Codex-compatible clients.

Or configure manually — add to ~/.claude/settings.json (global) or .claude/settings.json (per-project):

{
  "hooks": {
    "SessionStart": [{ "hooks": [{ "type": "command", "command": "ruby /path/to/hooks/session_start.rb", "timeout": 15000 }] }],
    "UserPromptSubmit": [{ "hooks": [{ "type": "command", "command": "ruby /path/to/hooks/saneprompt.rb" }] }],
    "PreToolUse": [{ "hooks": [{ "type": "command", "command": "ruby /path/to/hooks/sanetools.rb" }] }],
    "PostToolUse": [{ "hooks": [{ "type": "command", "command": "ruby /path/to/hooks/sanetrack.rb" }] }],
    "Stop": [{ "hooks": [{ "type": "command", "command": "ruby /path/to/hooks/sanestop.rb" }] }]
  }
}

Verify:

ruby scripts/hooks/saneprompt.rb --self-test
ruby scripts/hooks/sanetools.rb --self-test
ruby scripts/hooks/sanetrack.rb --self-test
ruby scripts/hooks/sanestop.rb --self-test

Codex Setup

Codex does not need Claude-style hook registration to use SaneProcess well.

  1. Keep the repo-level AGENTS.md accurate and practical.
  2. Put repo skills in .agents/skills/ so Codex can discover them.
  3. Keep the Codex registry and user skills coherent: ~/.codex/SKILLS_REGISTRY.md and ~/.codex/skills/.
  4. Configure MCP in ~/.codex/config.toml or project-scoped .codex/config.toml.
  5. Use the shared runtime guards for risky paths such as email writes.

Useful commands:

codex mcp list
codex mcp add context7 -- npx -y @upstash/context7-mcp
codex mcp add github -- npx -y @modelcontextprotocol/server-github

How It Works

SaneProcess has one SOP and two main runtime shapes:

  • Claude Code: 4 enforcement hooks plus 1 session-start bootstrap hook.
  • Codex: repo instructions, repo skills, MCP, SaneMaster.rb, and shared shell/script guards.

Claude's native hooks map to these lifecycle events:

Hook Event Purpose
session_start.rb SessionStart Kills orphans, resets state, bootstraps session
saneprompt.rb UserPromptSubmit Classifies intent, sets research requirements
sanetools.rb PreToolUse Blocks edits until research is done
sanetrack.rb PostToolUse Tracks failures, trips circuit breaker
sanestop.rb Stop Captures session summary

All state lives in a single HMAC-signed JSON file (.claude/state.json). File-locked for concurrent access. Tamper-detected via HMAC key (macOS Keychain or ~/.claude_hook_secret on Linux). Atomic writes via tempfile + rename.

Orphan Cleanup

On every session start, three cleanup passes run:

  1. Parent sessions — finds claude processes not in your current process tree
  2. MCP daemons — finds known MCP patterns (context7, apple-docs, xcodebuild, github, serena, central-memory, etc.) not in your session tree
  3. Subagents — finds claude --resume processes whose parent sessions are dead

Uses BFS process tree traversal. Your active session and any other active terminal sessions are never touched. The same process inspection logic also feeds the Codex MCP watchdog tooling.

Circuit Breaker

After tool execution, error signatures are normalized and tracked:

  • 3 consecutive failures → breaker trips
  • 3 identical error signatures → breaker trips

When tripped, all edit/write operations are blocked. Say reset breaker or rb- to clear after fixing the root cause.

Research Gate

Before any mutation (Edit, Write, Bash with side effects), research categories must be satisfied:

Category Satisfied By Required?
docs apple-docs MCP, context7 MCP Only if MCP configured
web WebSearch, WebFetch Always
github GitHub MCP Only if MCP configured
local Read, Grep, Glob Always

The gate adapts to your setup. With no configured MCPs, only web + local are required. Configure apple-docs/context7 and GitHub in .mcp.json to make docs and github required too. The installer shows recommended MCPs and install commands.

Tool Discovery Receipt

Before saying a tool is missing, using a workaround, or adding new tooling, generate the receipt:

ruby scripts/SaneMaster.rb tool_discovery --query "missing screenshot diff tool"

That receipt checks the skills registry, runs doctor, runs validation_report.rb, searches local scripts/docs, and writes proof artifacts to outputs/tool-discovery/.

Tool Categorization (Blast Radius)

Category Examples Blocked Until
Read-only Read, Grep, Glob, search Never
Local mutation Edit, Write Research complete
Sensitive files CI/CD, entitlements, Dockerfiles Confirmed per-file per-session
External mutation GitHub push Research complete

Security

  • HMAC-signed state — state.json is signed to detect tampering. Key stored in macOS Keychain (macOS) or ~/.claude_hook_secret with 600 permissions (Linux).
  • Blocked system paths — Prevents edits to /etc/, .ssh/, .aws/, .gnupg/
  • Inline script detection — python -c, ruby -e, node -e blocked as bash mutations
  • Sensitive file confirmation — First edit to CI/CD configs, entitlements, Dockerfiles requires confirmation
  • Fail-safe defaults — If a hook errors internally, it allows the operation (exit 0). Never blocks randomly.

Tests

452 tests across two frameworks:

Tier tests (178) — end-to-end enforcement scenarios:

Tier Count What
Easy 61 Basic functionality
Hard 58 Edge cases, state transitions
Villain 59 Adversarial bypass attempts

Self-tests (282) — per-hook unit tests:

Hook Tests
saneprompt 179
sanetools 44
sanetrack 32
sanestop 27 
# Run all tier tests
ruby scripts/hooks/test/tier_tests.rb

# Run per-hook self-tests
ruby scripts/hooks/saneprompt.rb --self-test

# Run a specific tier
ruby scripts/hooks/test/tier_tests.rb --tier villain

Configuration

Configurable via scripts/hooks/core/config.rb:

Setting Default What
Circuit breaker threshold 3 Consecutive failures before trip
File size warning 500 lines Yellow warning on edit
File size limit 800 lines Block the edit
Blocked paths /etc/, .ssh/, .aws/ System path protection

Troubleshooting

"BLOCKED: Research incomplete"

The hook is working correctly. Complete the required research categories before editing. The gate adapts — categories whose MCPs you don't have auto-skip. With no MCPs, only web + local are required.

Circuit breaker tripped

Say reset breaker or rb- in Claude after fixing the root cause.

central-memory MCP not working

SaneProcess includes a semantic memory MCP server backed by PostgreSQL + pgvector.

# Provision local database + extension + schema
cd ~/SaneApps/infra/SaneProcess/scripts/mcp-central-memory
./bootstrap-local.sh

# Verify MCP health
~/.codex/bin/check-mcps
codex mcp list | rg central-memory

Expected: check-mcps shows [PASS] central-memory.

If it fails:

  • Ensure OPENAI_API_KEY is available to the Codex app process.
  • Ensure PostgreSQL is running: brew services list | rg postgresql@17.

memory MCP search feels weak

Codex uses a separate graph-style memory MCP for entity/relation recall.

  • Config key: memory in /Users/sj/.codex/config.toml
  • Server: ~/SaneApps/infra/SaneProcess/scripts/mcp-memory-enhanced/server.mjs
  • Store: ~/.claude/memory/knowledge-graph.jsonl

Use memory when you need exact graph cross-reference like:

  • issue ↔ email thread ↔ root cause bucket
  • canonical entity names
  • linked neighbors from a matched node

Use central-memory when you need semantic recall across free-form notes.

Hooks not firing

Check that .claude/settings.json contains hook entries pointing to your scripts/hooks/ directory. Re-run init.sh if needed.

Requirements

  • macOS or Linux (process cleanup uses POSIX ps; HMAC key uses Keychain on macOS, file-based on Linux)
  • Ruby 3.x (brew install ruby on macOS; apt install ruby or dnf install ruby on Linux)
  • Claude Code (npm install -g @anthropic-ai/claude-code)
  • Codex (brew install codex, or the desktop/IDE client)

Included Skills

SaneProcess ships two reusable audit skills. Claude projects can install them into .claude/skills/. Codex-compatible projects should install them into .agents/skills/.

Critic (/critic)

6-persona adversarial code review. Spawns parallel subagents that each review from a different angle:

Agent Focus
Bug Hunter Logic errors, null safety, crashes
Data Flow Tracer Feature completeness across code paths
Integration Auditor Cross-system config consistency
Edge Case Explorer Unusual states, environments
UX Critic UI/UX quality, clarity, accessibility
Security Auditor Attack vectors, data exposure

Findings are deduplicated and ranked by severity. Issues caught by multiple agents get higher confidence scores.

Docs Audit (/docs-audit)

11-perspective documentation audit. Each perspective runs as its own subagent:

# Perspective Focus
1 Engineer Technical accuracy, API docs
2 Designer Screenshots, visual storytelling
3 Marketer Value proposition, "why should I care?"
4 User Advocate Onboarding clarity
5 QA Edge cases, gotchas, troubleshooting
6 Hygiene Duplication, terminology drift
7 Security Leaked secrets, PII, internal URLs
8 Freshness Stale examples, broken links
9 Completeness TODO placeholders, unchecked items
10 Ops Git hygiene, dependencies, certificates
11 Consistency Broken references in CLAUDE.md vs actual code

Produces a consolidated gap report with scores, critical issues, and recommended fixes — sorted by what the agent can fix vs what needs human action.

Skills are mirrored into .claude/skills/ and .agents/skills/ during init.sh setup.

Project Structure

scripts/
├── hooks/                    # All enforcement hooks
│   ├── session_start.rb      # SessionStart — orphan cleanup, state reset
│   ├── saneprompt.rb         # UserPromptSubmit — classify, set requirements
│   ├── sanetools.rb          # PreToolUse — block until research done
│   ├── sanetrack.rb          # PostToolUse — track failures, circuit breaker
│   ├── sanestop.rb           # Stop — session summary
│   ├── core/                 # Shared infrastructure
│   │   ├── config.rb         # Paths, thresholds, settings
│   │   ├── state_manager.rb  # Signed state file management
│   │   └── context_compact.rb
│   └── test/                 # Test suites
│       └── tier_tests.rb     # 178 enforcement tests
├── init.sh                   # Project installer
└── qa.rb                     # QA runner
AGENTS.md                     # Shared agent instructions
skills/
├── critic/                   # 6-persona adversarial code review
│   ├── SKILL.md
│   └── prompts/              # 6 agent prompts
└── docs-audit/               # 11-perspective documentation audit
    ├── SKILL.md
    └── prompts/              # 11 agent prompts
.claude/
├── rules/                    # Path-specific guidance (installed by init.sh)
│   ├── hooks.md              # Hook conventions
│   └── scripts.md            # Ruby script conventions
├── skills/                   # Installed by init.sh from skills/
└── settings.json             # Hook registration
.agents/
└── skills/                   # Codex-compatible repo skills (installed by init.sh)

Uninstall

Remove hook entries from .claude/settings.json. Delete scripts/hooks/. Delete .claude/state.json.

No global state modified. No daemons installed. No system changes.

License

MIT License. See LICENSE.

Built by SaneApps. Used in production across 7 projects.

Become a Contributor (Even if You Don't Code)

Are you tired of waiting on the dev to get around to fixing your problem?
Do you have a great idea that could help everyone in the community, but think you can't do anything about it because you're not a coder?

Good news: you actually can.

Copy and paste this into Claude or Codex, then describe your bug or idea:

I want to contribute to this repo, but I'm not a coder.

Repository:
https://github.com/sane-apps/SaneProcess

Bug or idea:
[Describe your bug or idea here in plain English]

Please do this for me:
1) Understand and reproduce the issue (or understand the feature request).
2) Make the smallest safe fix.
3) Open a pull request to https://github.com/sane-apps/SaneProcess
4) Give me the pull request link.
5) Open a GitHub issue in https://github.com/sane-apps/SaneProcess/issues that includes:
   - the pull request link
   - a short summary of what changed and why
6) Also give me the exact issue link.

Important:
- Keep it focused on this one issue/idea.
- Do not make unrelated changes.

If needed, you can also just email the pull request link to [email protected].

I review and test every pull request before merge.

If your PR is merged, I will publicly give you credit, and you'll have the satisfaction of knowing you helped ship a fix for everyone.

Yorumlar (0)

Sonuc bulunamadi