alhazen-skills

alhazen-skill-examples

Example skills for the Skillful Alhazen knowledge notebook framework — a TypeDB-powered scientific notebook for researchers building knowledge graphs from papers, notes, and domain data.

What is this?

This repo serves two purposes:

1. Claude Code plugin marketplace — install skills directly into Claude Code (v1.0.33+) without a full Skillful Alhazen project. Each skill is a self-contained plugin with its own dependencies.
2. Skillful Alhazen skills registry — add skills to skills-registry.yaml in a Skillful Alhazen project and make build-skills wires them in.

---

Skills

Skill	Plugin type	Description
jobhunt	self-contained	Track job applications, identify skill gaps, plan learning
alhazen-core	infrastructure	Starts TypeDB, loads base schema — required by multi-plugin installs
they-said-whaaa	standard	Credibility tracker — ingest YouTube + news, detect contradictions
scientific-literature	standard	Multi-source literature search (EPMC, PubMed, OpenAlex, bioRxiv)
alg-precision-therapeutics	standard	Rare disease mechanism investigation from a MONDO diagnosis
literature-trends	biomed	Trace hypothesis evolution across time windows in a literature cluster

---

Install via Claude Code Marketplace (recommended)

Requires Claude Code v1.0.33+. There are two plugin types with different install flows.

Self-contained plugins (zero-setup)

Self-contained plugins bundle everything they need — including the TypeDB init logic and base schema. TypeDB starts automatically on every session start via a SessionStart hook. No manual init required.

Currently available: jobhunt

/plugin marketplace add sciknow-io/alhazen-skill-examples
/plugin install jobhunt@alhazen-skills

That's it. On the next session start, the SessionStart hook runs:

uv run --project <plugin-root>/skills/jobhunt python <plugin-root>/skills/jobhunt/alhazen_core.py init

This starts the TypeDB Docker container (pulling the image if needed), creates the alhazen_notebook database, and loads both the base schema and jobhunt's domain schema. Subsequent session starts complete in under a second.

Expected output on first run:

{
  "success": true,
  "typedb": "running",
  "database": "alhazen_notebook",
  "database_created": true,
  "schema": "loaded",
  "extra_schema": "loaded",
  "message": "Alhazen core ready. Base schema and skill schema loaded."
}

Then use the skill:

uv run --project <skill-path> python <skill-path>/jobhunt.py list-pipeline

Standard plugins (install alhazen-core first)

Standard plugins depend on the alhazen-core infrastructure plugin. Install it first, then install domain skills individually.

Step 1 — Add the marketplace and install alhazen-core:

/plugin marketplace add sciknow-io/alhazen-skill-examples
/plugin install alhazen-core@alhazen-skills

Initialize the infrastructure (one-time):

/alhazen-core:init

Expected output:

{
  "success": true,
  "typedb": "running",
  "database": "alhazen_notebook",
  "schema": "loaded",
  "message": "Alhazen core ready."
}

Step 2 — Use a domain skill:

# Replace <skill-path> with your plugin cache path
# e.g. ~/.claude/plugins/cache/they-said-whaaa/
uv run --project <skill-path> python <skill-path>/they_said_whaaa.py list-figures

Marketplace structure

The repo-level catalog is at .claude-plugin/marketplace.json.

Self-contained plugin bundle (plugins/<name>/):

plugins/<name>/
  .claude-plugin/
    plugin.json             # Plugin metadata (no "requires": ["alhazen-core"])
  hooks/
    hooks.json              # SessionStart hook: runs alhazen_core.py init
  skills/<name>/
    SKILL.md                # Loaded at startup: triggers, quick start
    USAGE.md                # Full reference: commands, workflows, data model
    <name>.py               # CLI entry point
    alhazen_core.py         # Bundled copy of alhazen-core init logic
    alhazen_notebook.tql    # Bundled copy of the base schema
    schema.tql              # This skill's domain schema (auto-loaded by init)
    pyproject.toml          # uv dependency declaration

Standard skill (skills/<category>/<name>/):

skills/<category>/<name>/
  .claude-plugin/
    plugin.json             # Includes "requires": {"plugins": ["alhazen-core"]}
  SKILL.md / USAGE.md / skill.yaml
  <name>.py / pyproject.toml / schema.tql

---

Install via Skillful Alhazen (full project)

Add entries to skillful-alhazen/skills-registry.yaml: