Agent Caps

Define, validate, and audit agent capability manifests for safe skill installation

Agent Capability Manifest Toolkit

Validate, scaffold, and cross-check AI-agent capability manifests so agents stay swappable, auditable, and safe.

Keywords: agent, caps, security, manifest, cli, safety, python, open-source, automation, MIT

Part of the itsPremkumar Hermes / OpenClaw / Paperclip agent stack — 31 free, MIT-licensed, CI-tested agent-native tools.

What it does

Agents declare incompatible capabilities and dependencies, then break the system at install time. Agent Capability Manifest Toolkit solves this: Validate, scaffold, and cross-check AI-agent capability manifests so agents stay swappable, auditable, and safe.

Best for: Agent builders, platform engineers, and security reviewers shipping OpenClaw / Hermes / Paperclip agents.

Features

• Validate a manifest before publishing an agent to ClawHub
• Scaffold a standards-compliant manifest for a new agent
• Cross-check dependencies across a fleet of agents
• Gate CI on missing/invalid capability fields
• Audit third-party agents before granting machine access

Install

bash

# Requires Python 3.8+. No pip install needed.
curl -O https://raw.githubusercontent.com/itsPremkumar/agent-caps/main/agent_caps.py
# Or copy the file anywhere — it's self-contained.

Quick start

bash

python agent_caps.py self-test     # prove it works end-to-end
python agent_caps.py validate --help   # validate subcommand
python agent_caps.py scaffold --help   # scaffold subcommand
python agent_caps.py check-deps --help   # check-deps subcommand
python agent_caps.py schema --help   # schema subcommand

Use cases

1. Validate a manifest before publishing an agent to ClawHub
2. Scaffold a standards-compliant manifest for a new agent
3. Cross-check dependencies across a fleet of agents
4. Gate CI on missing/invalid capability fields
5. Audit third-party agents before granting machine access

Why choose this over alternatives

Alternative	Why this skill is better
Hand-written JSON schemas	agent-caps ships a single enforceable schema + validator, not ad-hoc docs.
Reading the manifest manually	It catches missing fields, bad versions, illegal statuses, and unknown deps automatically.
Trusting an agent's self-description	check-deps cross-verifies claims against reality before the agent goes live.

FAQ (SEO / AEO)

Q: What is an agent capability manifest?
A: A machine-readable JSON describing an agent's name, version, capabilities, dependencies, memory, tools, API, and status — the standard interface that lets agents be swapped or audited without breaking the host.

Q: Does agent-caps need internet or pip?
A: No. It is pure Python 3.8+ stdlib, zero dependencies, runs fully offline.

Q: Can I use it in CI?
A: Yes. Run validate as a CI gate; a non-zero exit blocks merges with malformed manifests.

Q: What does check-deps do?
A: Resolves declared dependencies against known agents/manifests and flags unknown or unmet requirements.

Geo / local reach

Built and maintained by @itsPremkumar (Chennai, India · serving developers worldwide). Free for individuals and teams everywhere. Documentation in English; tool output is locale-neutral.

CI integration

yaml

# .github/workflows/verify.yml
name: Verify
on: [push]
jobs:
  verify:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Self-test agent-caps
        run: python agent_caps.py self-test

Support

Free + MIT-0 (free, modifiable, no attribution required). Sponsor if useful:

• GitHub Sponsors: https://github.com/sponsors/itsPremkumar
• Buy Me a Coffee: https://buymeacoffee.com/itsPremkumar

⭐ Star on GitHub