Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/docxology/codomyrmexnpx agentmods add skills/docxology/codomyrmex/paiWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/docxology/codomyrmex/pai)<a href="https://agentmods.dev/skills/docxology/codomyrmex/pai"><img src="https://agentmods.dev/badge/skills/docxology/codomyrmex/pai/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/docxology/codomyrmex/pai"><img src="https://agentmods.dev/badge/skills/docxology/codomyrmex/pai.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00185 | $0.02022 |
| Opus 5.5 | $0.00074 | $0.00809 |
| Sonnet 5.5 | $0.00037 | $0.00404 |
| Haiku 4.5 | $0.00018 | $0.00202 |
Grade A, and why
Codomyrmex scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 13d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Codomyrmex Skill for PAI
Version: v1.3.0 | Type: Infrastructure Skill | MCP: codomyrmex-mcp-server | Skills: 15 | Tools: 627 (@mcp_tool; see docs/reference/inventory.md)
Canonical copy for install-from-repo: SKILL.md at repository root. This file stays aligned with that manifest for PAI pack resolution.
Description
Full-spectrum coding workspace skill providing 627 production @mcp_tool lines across 130 top-level modules for AI-assisted development, code analysis, testing, documentation generation, and workflow automation.
Installation
# codomyrmex is not published on PyPI; install it from a checkout
uv sync
Quick Start
# MCP server (for PAI agent consumption)
from codomyrmex.agents.pai.mcp_bridge import create_codomyrmex_mcp_server
server = create_codomyrmex_mcp_server()
server.run()
# Direct Python calls (no MCP overhead)
from codomyrmex.agents.pai import verify_capabilities, trust_all, trusted_call_tool
verify_capabilities() # Step 1: Audit & promote safe tools
trust_all() # Step 2: Grant full execution trust
result = trusted_call_tool("codomyrmex.write_file", path="x.py", content="...")
If confirmation mode is enabled for destructive tools, the first call returns a one-use token. Re-submit that token with the same validated arguments; changing the payload invalidates the confirmation rather than authorizing a different action.
Live tool and prompt inventory
Static documentation cannot list every MCP tool (the surface is large and changes with the tree). Use:
from codomyrmex.agents.pai import get_skill_manifest
m = get_skill_manifest()
len(m["tools"]) # merged static + dynamic tool count
m["prompts"] # prompt names and descriptions
m["workflows"] # bundled workflow definitions
Refresh documented decorator and file counts after MCP changes: uv run python scripts/doc_inventory.py (optional: --manifest for runtime tool count).
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- __init__.py 3.5 KB runs code
- _models.py 7.8 KB runs code
- _modules.py 8.7 KB runs code
- _systems.py 7.6 KB runs code
- _verification.py 5.1 KB runs code
- AGENTS.md 1.8 KB
- mcp_bridge.py 1.5 KB runs code
- MCP_TOOL_SPECIFICATION.md 7.4 KB
- mcp/__init__.py 40 B runs code
- mcp/AGENTS.md 1.5 KB
- mcp/definitions.py 14 KB runs code
- mcp/discovery.py 11 KB runs code
- mcp/path_policy.py 3.7 KB runs code
- mcp/proxy_tools.py 14 KB runs code
- mcp/README.md 520 B
- mcp/server.py 21 KB runs code
- mcp/SPEC.md 5.5 KB
- mcp/trust_metadata.py 3.5 KB runs code
- pai_bridge.py 2.7 KB runs code
- pai_client.py 4.8 KB runs code
- pai_webhook.py 7.1 KB runs code
- PAI.md 4.1 KB
- pm/.env.example 836 B
- pm/.gitignore 19 B
- pm/AddTask.ts 14 KB runs code
- pm/AGENTS.md 2.7 KB
- pm/bun.lock 1.3 KB
- pm/CompleteProject.ts 15 KB runs code
- pm/config.ts 2.6 KB runs code
- pm/CreateMission.ts 8.3 KB runs code
- pm/CreateProject.ts 11 KB runs code
- pm/DataModels.ts 2.9 KB runs code
- pm/DeleteMission.ts 7.0 KB runs code
- pm/DeleteProject.ts 7.7 KB runs code
- pm/GitHubSync.ts 62 KB runs code
- pm/helpers.ts 2.6 KB runs code
- pm/ListMissions.ts 9.7 KB runs code
- pm/ListProjects.ts 9.6 KB runs code
- pm/ListTasks.ts 11 KB runs code
- pm/MissionDashboard.ts 17 KB runs code
- pm/package.json 909 B
- pm/PAI.md 2.5 KB
- pm/PMDashboard.ts 37 KB runs code
- pm/ProjectDashboard.ts 16 KB runs code
- pm/README.md 1.1 KB
- pm/routes/AGENTS.md 1.7 KB
- pm/routes/awareness.ts 8.9 KB runs code
- pm/routes/calendar.ts 8.7 KB runs code
- pm/routes/dispatch.ts 12 KB runs code
- pm/routes/email.ts 42 KB runs code
- pm/routes/github.ts 5.1 KB runs code
- pm/routes/interview.ts 7.3 KB runs code
- pm/routes/missions.ts 2.6 KB runs code
- pm/routes/projects.ts 3.5 KB runs code
- pm/routes/README.md 627 B
- pm/routes/SPEC.md 883 B
- pm/routes/tasks.ts 2.2 KB runs code
- pm/server.ts 4.9 KB runs code
- pm/services/AGENTS.md 1.3 KB
- pm/services/oauth.ts 7.7 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 13d ago Changed · -1 lines 98e82056ed11
- 2mo ago First seen · 161 lines · 185 tokens per session scan A aec22ab2f09c
Codomyrmex is a skill published in the GitHub repository docxology/codomyrmex (11 stars, last pushed yesterday), licensed MIT. It adds 185 tokens to every session and 2,022 once invoked, about $0.0007 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Codomyrmex compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| Codomyrmexdocxology/codomyrmex | Skill | 11 | yesterday | MIT | A | 185 | Claude Code, Codex |
| neurolink-guidejuspay/neurolink | Skill | 169 | today | MIT | A | 65 | Claude Code |
| github-commentingjuspay/neurolink | Skill | 169 | today | MIT | A | 41 | Claude Code, Codex |
| repo-conventionsjuspay/neurolink | Skill | 169 | today | MIT | A | 39 | Claude Code, Codex |
| tautullijmagar/claude-homelab | Skill | 88 | 1mo ago | MIT | A | 74 | Claude Code |
| notebooklmjmagar/claude-homelab | Skill | 88 | 1mo ago | MIT | A | 67 | Claude Code |
| gh-address-commentsjmagar/claude-homelab | Skill | 88 | 1mo ago | MIT | A | 85 | Claude Code, Codex |