Skip to content

cogni-dao/cogni / .claude/skills/node-catalog

node-catalog

Verified

The node roster + operator-governance map for devs working in this repo. Use to answer "what nodes exist / are they live / what are their URLs", "who owns a node and what can the operator do to it", or "which parts are the node's vs the operator's". A where-to-look reference, NOT a doer — for actions use deploy-node / node-wizard-expert / promote / manage-node-envs. Triggers: "list the nodes", "node roster", "node URLs", "is <node> live", "operator's role/permissions over nodes".

node-catalog — node roster + operator-governance map

The roster is live state — read it, never hardcode

WantSource
Who's registered + owner (per env)GET /api/v1/nodes (owner Bearer) — the Postgres nodes SSOT. Per-env status: GET /api/v1/nodes/{id}/deploy-state.
Git-declared in-repo nodesinfra/catalog/*.yaml type:node — read the rows, do not trust any list written here (litellm/openfga=infra, scheduler-worker=service).
Is it actually servingcurl https://<host>/version from outside the cluster (buildSha, not workflow-green).

URL rule (verify-buildsha.sh): operator = the bare env domain; every other node = <node>-<envprefix>.<base>.

  • prod cognidao.org / <node>.cognidao.org · preview preview.cognidao.org / <node>-preview… · candidate-a test.cognidao.org / <node>-test…

External submodule nodes (e.g. blue, oss, habitat) carry their own repo-spec node_id — they show in GET /api/v1/nodes but not the in-repo catalog. Not-in-catalog + not-serving = an external node's own deploy, not an in-repo gap.

Operator's role: node declares shape; operator wires environment

  • Node owns (edit in the node repo): app code, packages, base manifests, image build, declarations (repo-spec.yaml, DB schema, secrets-catalog.yaml).
  • Operator owns (the deploy plane): the catalog, provisioning, DNS, the deploy verb POST /api/v1/deploy/workloads and its sequencer, secret values, env ownership. Nodes declare .cogni/workloads/<name>.yaml; see cicd-e2e-required-sequence.
  • Identity SSOT is one value: repo-spec node_id = nodes.id = OpenFGA node:<id> = Loki node label. Address a node by that id (slug also resolves). The operator is itself a uniform node.

What the operator can do to a node — and its limits (all OpenFGA-gated)

Every write is (node_id, env)-scoped, resolves the node via the registry, gates OpenFGA node:<id>, and runs as the operator's own identity — the caller holds only an API key.

ActionEndpointRole
Flight → candidate-aPOST /api/v1/vcs/flightdeveloper
Merge a node PR (on green)POST /api/v1/vcs/mergedeveloper
Promote → productionPOST /api/v1/deploy/promoteproduction_promoter
Write/rotate a node secret valuePOST /api/v1/nodes/{id}/secretssecrets_manager
Read deploy-state / logsGET /api/v1/nodes/{id}/{deploy-state,observability/logs}developer

owner (the repo-spec governance-approver wallet) RLS-owns the row and approves access-requests (POST /nodes/{id}/access-requests → POST /nodes/{id}/developers). 403 authz_denied = no grant; 503 authz_unavailable = OpenFGA store unbootstrapped (not a denial).

Limits: the operator is the deploy plane only — code, work-items, and knowledge live elsewhere (node repo · work API · the Dolt hub). It does not edit a node's app code or run/freeze a node's own CI (external nodes keep their GitHub Actions), and it can't promote/secret a node absent from that env's registry (node_not_found → seed first).

Canon (verify against these; don't restate)

writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 319kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 118kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall