library

HUAOMP Library — acquire, catalog, query, and curate knowledge across 20 departments. Receipt-backed acquisitions, SHA-256 dedup, three-tier access classification.

SKILL.md

GitHub

library

HUAOMP Library — the HUMMBL knowledge repository. Catalogs, stores, and queries knowledge across 20 departments with receipt-backed acquisitions and SHA-256 content dedup.

When to use

Explicit invocation

  • [library] acquire <path> — register a document into the library
  • [library] acquire-batch <dir> — register all documents in a directory
  • [library] query <text> — search the catalog by text
  • [library] query --department <dept> — filter by department
  • [library] query --tier <T1-T13> — filter by knowledge tier
  • [library] query --lenses H,O — filter by HUAOMP lenses
  • [library] stats — collection statistics
  • [library] departments — per-department breakdown
  • [library] ingest-ledger <path> — import loop-ledger JSONL entries

Auto-propose

  • When a research session produces documents that should be cataloged
  • When importing external sources (Gutenberg, arxiv, Wikipedia)
  • When querying the fleet's knowledge base for prior research

Departments (20)

Department Wing Name Source
canonical_foundations The Roots T1
empirical_research The Observatory T2
applied_systems The Workshop T3
agentic_systems The Hive T4
engineering_practice The Forge T5
governance The Court T6
emerging_technology The Horizon T7
cognition The Mind T8
economics The Exchange T9
collaboration The Forum T10
security The Citadel T11
complexity The Lattice T12
reasoning The Compass T13
arcana The Pantheon Fleet
operations The Bridge Fleet
skills The Library of Skills Fleet
rules The Codex Fleet
external_humanities The Alexandria Wing Gutenberg
external_science The Tigris Wing arxiv
external_world_knowledge The Bosphorus Wing Wikipedia

Access tiers

  • public — safe to share externally (default)
  • internal — fleet-only (AARs, operations, bus history)
  • restricted — operator-only (tokens, secrets, credentials)

Storage

  • Catalog: ~/.agents/library/catalog.jsonl (append-only JSONL)
  • Index: ~/.agents/library/catalog_index.json (item_id → line, sha256)
  • Receipts: via hummbl-governance kernel ReceiptEngine

Execution

acquire

from hummbl_governance.library import Library, Department

lib = Library(state_dir=Path("~/.agents/library"))
entry = lib.acquire(Path("docs/research/findings.md"), agent_id="devin")

query

results = lib.query(text="governance", department=Department.GOVERNANCE)

ingest-ledger

from hummbl_governance.library import import_loop_ledger
successes, failures = import_loop_ledger(
    Path("~/.agents/loop-ledger/intel-surge-ledger.jsonl"),
    lib,
)

Dependencies

  • hummbl_governance.kernel.ReceiptEngine (optional — for receipt-backed acquisitions)
  • hummbl_governance._file_lock (for concurrent write safety)
  • Python stdlib only — no third-party dependencies