An open API service indexing awesome lists of open source software.

https://github.com/hummbl-dev/hummbl-skills

HUMMBL Agent Skills — canonical skill definitions, CI-validated SKILL.md files, skill creation templates, and cross-agent skill export tooling. Governed replacement for ad-hoc skill collections.
https://github.com/hummbl-dev/hummbl-skills

Last synced: 30 days ago
JSON representation

HUMMBL Agent Skills — canonical skill definitions, CI-validated SKILL.md files, skill creation templates, and cross-agent skill export tooling. Governed replacement for ad-hoc skill collections.

Awesome Lists containing this project

README

          

# HUMMBL Agent Skills

*Canonical skill definitions, CI-validated SKILL.md files, and cross-agent skill export tooling. Maintained by [HUMMBL Research Institute](https://hummbl.io).*

## What this is

A governed, CI-validated catalog of agent skills for the HUMMBL multi-agent platform. Replaces ad-hoc skill collections with structured, versioned, testable skill definitions.

## Structure

```
skills/ # Canonical skill definitions
aar/ # After Action Report generation
sitrep/ # Situational report
triage/ # Adopt/Adapt/Avoid evaluation
...
template/ # Skill creation templates
SKILL.md # Canonical SKILL.md template
spec/ # Skill specification and validation rules
schema.json # JSON Schema for SKILL.md validation
tools/ # Skill management tooling
validate.py # Validate all SKILL.md files against spec
export.py # Export skills to other agent formats
score.py # Score skill quality/completeness
```

## Skill Format

Every SKILL.md must include:

- **Frontmatter**: name, description, triggers, chains, base120 mapping
- **Workflow**: step-by-step instructions for the agent
- **Constraints**: what the agent must not do
- **Evidence**: required receipts/outputs

See `template/SKILL.md` for the canonical template.

## Principles

- **CI-validated** — every skill passes schema validation before merge
- **Agent-agnostic** — skills export to Claude Code, Codex, OpenClaw, and custom agents
- **Versioned** — skill changes are tracked with SemVer
- **Governed** — follows HUMMBL append-only audit standards
- **Stdlib-only tooling** — validation and export tools use Python stdlib only

## Status

**v0.1 — BOOTSTRAP COMPLETE.** The repository now contains:

- `skills/`, `template/`, `spec/`, and `tools/` directories
- canonical `template/SKILL.md`
- schema-backed validation in `spec/schema.json` and `tools/validate.py`
- cross-agent export tooling in `tools/export.py`
- three initial skills: `data-export`, `mtsmu-review`, and `tag-release`
- CI validation, scoring, self-review, export dry-run, and golden export comparison

Current bootstrap receipts:

```text
python tools/validate.py # 3 skills pass
python tools/score.py # average score: 100%
python tools/self_review.py # 0 issues, 0 warnings
python tools/export.py --dry-run
```

## Related

- [claude-config](https://github.com/hummbl-dev/claude-config) — production Claude Code configuration (rules, skills, hooks, memory)
- [hummbl-governance](https://github.com/hummbl-dev/hummbl-governance) — governance runtime
- [base120](https://github.com/hummbl-dev/base120) — Base120 mental model reference

## License

Apache License 2.0.