Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 13 additions & 5 deletions CLAUDE.md
Original file line number Diff line number Diff line change
Expand Up @@ -80,8 +80,9 @@ The suite has twelve Mechs, all under https://github.com/CultureBotAI, plus the
orchestrator `culturebotai-claw`. Three have pages here; the others link to their
own GitHub Pages sites. DUFMech's card links its Pages dashboard and seed
snapshot, the source of its count; the dashboard went live on 2026-10-05.
CLAW's manifest lists eleven members, including PathwayMech; DUFMech is a
separately listed suite member not yet declared in that manifest.
CLAW membership is read from `_fleet/data/manifest.json`, independently of
suite coverage. An absent declaration remains undeclared; neither a suite card
nor a website feature verdict implies admission.

`mechs.md` (`/mechs/`) is **generated**. Edit its sources, never the page:
`_fleet/mechs_template.md` (prose, cards, the cross-reference list),
Expand Down Expand Up @@ -110,7 +111,8 @@ Sources, refreshed differently:
`_fleet/data/manifest.json`, refreshed with
`scripts/fleet/refresh_manifest.py --claw-root <claw checkout>`, separately
from the census. Its tooltips quote claw's reasons verbatim. The site lists
twelve Mechs, while this manifest currently lists eleven.
twelve Mechs; the membership count and undeclared-project prose are derived
from the snapshot, not maintained as a separate count.
- **All twelve corpora:** PathwayMech records and DUFMech seed families use the
common census, subset, statistics and site-audit pipeline. DUFMech's frozen
JSON worklist is validated and projected into individual family documents;
Expand Down Expand Up @@ -138,14 +140,20 @@ Sources, refreshed differently:
hand-curated too, but judged on the live sites on its `checked_on` date rather
than at the pins: one verdict per feature per site, with criteria in its
catalogue and each site's github-pages `deployed_revision`.
A later site-only recheck adds that Mech's own `checked_on` and `scope`
together, backed by a successful deployed revision and browser evidence for
every feature. Its date and method render separately; leave the baseline
date and other sites unchanged. Schema-documentation prose reflects the
website verdict, never an inference about whether a repository has a schema.

Last full data refresh was 2026-10-05, pinned at 03:20 UTC. All twelve Mechs
participate in the census, overlap assets and repository statistics, including
PathwayMech's 152 records and DUFMech's 6,532 seed families. TraitMech has 1,005
records and CellStructureMech 885 at those pins. TaxonMech joined the census in
#87. Shared identifiers draw vocabulary chords; direct-reference arrows require
committed evidence as described in `_fleet/README.md`. CLAW still declares only
eleven members, so DUFMech's capability row remains explicitly undeclared.
committed evidence as described in `_fleet/README.md`. Membership and website
audits can advance independently of these census pins; neither refresh is
permission to redate or recount the other inputs.

The suite licensing policy is CC BY 4.0 for data and BSD-3-Clause for code.
Redistributed source material retains its applicable terms.
Expand Down
33 changes: 24 additions & 9 deletions _fleet/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -24,19 +24,34 @@
the deployment ID is `success` before recording its SHA; a queued or failed
deployment request does not identify the currently served site. The update
checker reports such unverified states as `NOT CHECKED`.
For a later site-only recheck, add `checked_on` and `scope` together to that
Mech's entry, update its `deployed_revision` to the successfully served SHA,
and reassess every feature against browser evidence. The row date must be
a canonical `YYYY-MM-DD` no earlier than the baseline. The assembler shows
the exception date above the table and its method and revision with that
site's evidence. The baseline `scope` is also rendered, clearly labeled.
Leave top-level `checked_on`, other sites' entries and census pins unchanged;
no per-site fields means the baseline date and method still apply. A partial
feature spot-check cannot redate the entire row. An entry without a published
site cannot carry a website recheck. Adding these optional fields is a
provenance assertion, not an automatic browser test or approval of a feature.
- `audit_notes.json` — hand-checked explanations of each source and its counted
entities. PathwayMech records and DUFMech seed families use the same pinned
census, statistics and site-audit pipeline as the other ten Mechs.
- Other `data/` files — derived numbers: `prefix_census.json`, `subsets_summary.json`, `cells_summary.json`, `fleet_data.json`, `mech_stats.json`; `site_audit.json` is the provenance record, written by `scripts/fleet/build_site_audit.py` from the snapshot, the live sites and the audited notes in `_fleet/audit_notes.json`.

## Membership and capability updates

The site currently lists twelve Mechs. CLAW's manifest lists eleven, including
PathwayMech; DUFMech is listed separately without a manifest membership declaration. The
October 5 vocabulary census, shared-term assets and repository statistics cover
all twelve Mechs. DUFMech's contribution is a measured seed worklist, labeled
as families rather than curated mechanism records. Its capability row reports
undeclared values without inventing CLAW membership.
The site lists twelve Mechs. Membership counts, badges and prose about
undeclared projects come from `data/manifest.json`; suite coverage does not
imply CLAW admission. A project's capability row remains undeclared until the
snapshot contains its declaration. Tests exercise both an eleven-member
fixture without DUFMech and a twelve-member fixture with synthetic DUFMech
declarations, independently of the current snapshot. The October 5 vocabulary
census, shared-term assets and repository statistics cover all twelve Mechs.
DUFMech's contribution to that census is a measured seed worklist, labeled as
families rather than curated mechanism records. Its schema-documentation prose
uses the website feature verdict, not membership or an inferred schema absence.

Use a CLAW checkout at the desired published `main` revision. The refresh reads
committed Git blobs at that checkout's HEAD; uncommitted changes are excluded.
Expand Down Expand Up @@ -396,9 +411,9 @@ status. A Mech without that field still receives its merged-PR count and pinned
provenance, rather than an invented zero reviewed count. All twelve repositories
participate in statistics collection.

CLAW's ad3862e snapshot declares eleven members and enables PathwayMech's KGX and
SSSOM exporters. DUFMech remains outside that upstream manifest; its capability
row is explicitly undeclared. The capability table preserves CLAW's full
CLAW's ad3862e snapshot at this refresh declared eleven members and enabled
PathwayMech's KGX and SSSOM exporters. DUFMech was outside that upstream manifest;
its capability row was explicitly undeclared. The capability table preserves CLAW's full
catalogue and recorded reasons. Capability adoption does not establish that all
CLAW workflows execute unattended.

Expand Down
2 changes: 1 addition & 1 deletion _fleet/audit_notes.json
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,6 @@
"TraitMech": "1,005 data/traits records at the pin, equal to the landing tile: 427 REVIEWED, 490 PROPOSED, 50 DEPRECATED and 38 SEEDED. The README and corpus contain 718 records with causal graphs. The page shows 10 populated categories and 47.5% embedded (477 of 1,005).",
"PathwayMech": "152 data/pathways records at the pin and 152 links in the published record browser. They contain 4,690 mechanistic edges and 15 taxa. Published record pages now display labeled edges, evidence quotes and linked references. Current committed outputs include KGX nodes/edges and SSSOM source mappings. The optional curation_history schema retains its explicit TaxonMech CurationEvent reference; no record currently populates it. PathwayMech is measured by the common census, subset and statistics pipeline.",
"DUFMech": "6,532 seed families in data/worklists/interpro-pfam-duf-2026-10-01.json, validated against its frozen manifest: 4,533 UNKNOWN_CANDIDATE and 1,999 KNOWN_HISTORICAL_DUF. Each JSON row is one measured family; the TSV is a redundant serialization. Counts, heatmap occurrences and shared-term links measure this seed worklist without claiming curated mechanism records. Pfam and InterPro accessions are structured family identifiers. The repository includes a generated dashboard with all 6,532 families marked UNSCORED, but the checked public Pages URLs return 404. CLAW membership remains undeclared.",
"culturebotai-claw": "The ad3862e snapshot declares 11 canonical fleet members, including PathwayMech, 20 governed artifacts and 26 capabilities per member. PathwayMech KGX and SSSOM export contracts are enabled. DUFMech is included in the website census and statistics but remains undeclared in the upstream manifest; its capability row must not invent adoption states."
"culturebotai-claw": "The published 73958ed snapshot declares 12 canonical fleet members, including DUFMech, 20 governed artifacts and 26 capabilities per member. This manifest-only refresh retains the October 5 Mech census, card and statistics pins; the 24 website-feature criteria are a separate evidence surface. DUFMech has 20 enabled, 3 disabled and 3 not_applicable canonical declarations, with reasons and settings preserved verbatim. Its enabled offline adapters include genuine OAK correspondence to frozen Pfam labels, retained-abstract knowledge-gap proposals and KGX/SSSOM exports. Knowledge-gap scheduled execution remains false; provider-backed deep research, METPO proposals and causal-graph coverage remain disabled. This governance-pin update does not claim scientific characterization, live source freshness or a full data refresh."
}
}
187 changes: 185 additions & 2 deletions _fleet/data/manifest.json
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,8 @@
"version": 1,
"source": {
"repository": "CultureBotAI/culturebotai-claw",
"revision": "ad3862e7a5bd771f94c320bde5fd364d3460884d",
"url": "https://github.com/CultureBotAI/culturebotai-claw/blob/ad3862e7a5bd771f94c320bde5fd364d3460884d/src/kg_microbe_fleet/fleet.yaml"
"revision": "73958edb97994bc16c5404443956ee64c47bfe9b",
"url": "https://github.com/CultureBotAI/culturebotai-claw/blob/73958edb97994bc16c5404443956ee64c47bfe9b/src/kg_microbe_fleet/fleet.yaml"
},
"capability_catalogue": {
"id_label_validation": {},
Expand All @@ -28,6 +28,9 @@
},
"knowledge_gap_scan": {
"settings": {
"scheduled": {
"type": "boolean"
},
"window": {
"type": "integer",
"required_when_enabled": true,
Expand Down Expand Up @@ -2014,6 +2017,186 @@
"reason": "Every record is one causal graph: CURIE nodes in the record-level taxa, participants and reactions lists plus the record itself, and one record-level mechanistic_edges list. That is neither graph_list nor node_nested_edges, so the coverage reader would report no record covered. Enable once the reader has a record-level shape."
}
}
},
"DUFMech": {
"key": "dufmech",
"github": "CultureBotAI/DUFMech",
"capabilities": {
"id_label_validation": {
"status": "enabled",
"reason": "conf/id_label_targets.yaml uses genuine OAK against a verified frozen Pfam reference, captured independently from family record labels by src/dufmech/id_labels.py. This checks retained-source correspondence, not live ontology freshness or scientific characterization.",
"reason_claims": {
"present": [
"conf/id_label_targets.yaml",
"src/dufmech/id_labels.py"
]
}
},
"curation_history": {
"status": "enabled"
},
"strict_validation": {
"status": "enabled"
},
"schema_sync": {
"status": "enabled"
},
"release_management": {
"status": "enabled"
},
"build_coordination": {
"status": "enabled"
},
"documentation": {
"status": "enabled"
},
"testing": {
"status": "enabled"
},
"refactoring": {
"status": "enabled"
},
"coordination_hooks": {
"status": "enabled"
},
"deep_research": {
"status": "disabled",
"reason": "conf/deep_research_provider.yaml and src/dufmech/research.py support offline planning and retained simulated results only. Provider execution is refused; enable after reviewed provider-backed adoption.",
"reason_claims": {
"present": [
"conf/deep_research_provider.yaml",
"src/dufmech/research.py"
]
}
},
"edison_key_discovery": {
"status": "not_applicable",
"reason": "DUFMech has no Edison research-provider integration."
},
"environment_coverage": {
"status": "not_applicable",
"reason": "DUF/Pfam family evidence is not a culture-media, ingredient or microbial-community environment inventory."
},
"knowledge_gap_scan": {
"status": "enabled",
"reason": "conf/kgscan_config.yaml and src/dufmech/knowledge_gaps.py use shared sentence scoring over retained abstracts and emit timestamped review packets. Explicit digest-bound review is required before native overlay acceptance. This offline adapter is not a live scheduled scan.",
"reason_claims": {
"present": [
"conf/kgscan_config.yaml",
"src/dufmech/knowledge_gaps.py"
]
},
"settings": {
"window": 25,
"scheduled": false
}
},
"vendored_sync": {
"status": "enabled"
},
"corpus_statistics": {
"status": "enabled",
"settings": {
"fields": [
"id",
"name",
"seed_status",
"characterization_status",
"curation_status",
"provenance"
]
}
},
"page_budgets": {
"status": "enabled",
"settings": {
"site_path": "pages",
"budgets_path": "conf/pages_budgets.json"
}
},
"kgx_export": {
"status": "enabled",
"settings": {
"nodes_path": "exports/kgx/nodes.tsv",
"edges_path": "exports/kgx/edges.tsv"
}
},
"writer_audit": {
"status": "enabled",
"settings": {
"search_dirs": [
"src/dufmech",
"scripts"
],
"save_helpers": [
"write_records",
"new_history"
],
"validators": [
"validate_record\\(",
"_validate_record\\(",
"_validate_content\\("
],
"curation_markers": [
"new_history\\("
]
}
},
"sssom_export": {
"status": "enabled",
"settings": {
"mapping_globs": [
"exports/sssom/*.sssom.tsv"
]
}
},
"metpo_proposal": {
"status": "disabled",
"reason": "No proposals/ directory exists. Enable when family mechanisms need new METPO terms and the repository adopts the proposal contract.",
"reason_claims": {
"absent": [
"proposals"
]
}
},
"source_queue": {
"status": "enabled",
"settings": {
"queue_path": "curation/source_queue.tsv",
"extensions": [
"implementation",
"reviewed_on",
"review_basis",
"script",
"artifacts",
"license_url"
],
"required_when_adopted": [
"script",
"artifacts",
"license_url",
"review_basis"
]
}
},
"site_contract": {
"status": "enabled",
"settings": {
"site_path": "pages"
}
},
"source_catalogue": {
"status": "enabled"
},
"unmapped_inventory_input": {
"status": "not_applicable",
"reason": "Protein-family evidence is not a culture-medium ingredient corpus."
},
"causal_graph_coverage": {
"status": "disabled",
"reason": "The schema supports scientific assertions, but the seeded corpus has no curated mechanistic assertions or adopted causal-graph projection in a supported shared coverage shape."
}
}
}
},
"artifact_count": 20
Expand Down
8 changes: 4 additions & 4 deletions _fleet/data/site_audit.json
Original file line number Diff line number Diff line change
Expand Up @@ -45,10 +45,10 @@
},
{
"repo": "culturebotai-claw",
"sha": "ad3862e7a5bd771f94c320bde5fd364d3460884d",
"commit_date": "2026-10-03T10:49:13Z",
"readme_url": "https://github.com/CultureBotAI/culturebotai-claw/blob/ad3862e7a5bd771f94c320bde5fd364d3460884d/README.md",
"notes": "The ad3862e snapshot declares 11 canonical fleet members, including PathwayMech, 20 governed artifacts and 26 capabilities per member. PathwayMech KGX and SSSOM export contracts are enabled. DUFMech is included in the website census and statistics but remains undeclared in the upstream manifest; its capability row must not invent adoption states."
"sha": "73958edb97994bc16c5404443956ee64c47bfe9b",
"commit_date": "2026-10-07T20:46:26Z",
"readme_url": "https://github.com/CultureBotAI/culturebotai-claw/blob/73958edb97994bc16c5404443956ee64c47bfe9b/README.md",
"notes": "The published 73958ed snapshot declares 12 canonical fleet members, including DUFMech, 20 governed artifacts and 26 capabilities per member. This manifest-only refresh retains the October 5 Mech census, card and statistics pins; the 24 website-feature criteria are a separate evidence surface. DUFMech has 20 enabled, 3 disabled and 3 not_applicable canonical declarations, with reasons and settings preserved verbatim. Its enabled offline adapters include genuine OAK correspondence to frozen Pfam labels, retained-abstract knowledge-gap proposals and KGX/SSSOM exports. Knowledge-gap scheduled execution remains false; provider-backed deep research, METPO proposals and causal-graph coverage remain disabled. This governance-pin update does not claim scientific characterization, live source freshness or a full data refresh."
},
{
"repo": "CultureMech",
Expand Down
Loading
Loading