菜单
Discover Clay actions and manage the local action catalog - use when searching for an available action/integration, checking catalog coverage against the strict proof registry, classifying action safety, or generating action templates.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Set up auth, session, and profile config for clay-cli before running any command - use when a Clay command fails with "session expired", when starting fresh in a new checkout, or when configuring workspace/folder/workbook IDs for the first time.
Extract Clay table/workbook data to local files - use when you need a full-fidelity workbook export, a template-shaped snapshot of a table, or raw paginated row dumps for offline processing.
Run a full Clay playbook end to end - intake, plan, prepare, sample, verify, scale - use when a plain-language request should be routed to one of the 12 packaged playbooks and operated through to a safe scale decision, not just a single ad-hoc command.
Build Clay list-building sources - company-to-people waterfalls, source preview/import, and webhook intake - use when adding a new list-building source to a table or converting a companies table into a people search.
Build and modify Clay tables, views, fields, and records with clay-v2.js - use when creating a new table/workbook, adding or editing columns (text/formula/http-api/use-ai/action), managing views, or writing declarative table specs.
Read-only exploration of Clay tables, workbooks, sources, and workspace metadata - use when you need to understand a table's shape, check run status, list workspaces/tables, or pull credit usage without changing anything.
| name | clay-browse-actions |
| description | Discover Clay actions and manage the local action catalog - use when searching for an available action/integration, checking catalog coverage against the strict proof registry, classifying action safety, or generating action templates. |
These commands work against a locally-held action catalog (a JSON dump of Clay's
action types) plus, for some, an integration-library/registry.yaml that curates
which actions are "proven safe" to reuse. Most are offline once you have a catalog
file; discovery against live Clay needs a session.
node clay-v2.js search-actions "email finder" --workspace {{workspace_id}}
node clay-v2.js actions-catalog --workspace {{workspace_id}} --query "linkedin"
search-actions defaults to searching across 12 action types; narrow with
--types t1,t2.actions-catalog --report file writes the full catalog dump to a file for the
offline commands below to consume.node clay-v2.js normalize-actions-catalog actions-catalog.raw.json --out actions-catalog.normalized.json
node clay-v2.js catalog-delta stored-catalog.json new-catalog.json --out delta-report.json
# equivalent standalone form:
node lib/catalog-delta.js stored-catalog.json new-catalog.json --out delta-report.json
node lib/catalog-coverage-dashboard.js --raw actions-catalog.raw.json --strict-registry integration-library/registry.yaml --md coverage.md --limit 25
node lib/safety-classifier.js --input actions-catalog.raw.json --output classifications.json --per-action
node lib/proof-strategy-classifier.js actions-catalog.raw.json proof-strategies.json
node lib/action-template-generator.js actions-catalog.raw.json --query "email" --limit 10 --format yaml --out templates.yaml
node lib/blocked-action-report.js --catalog actions-catalog.raw.json --json-out blocked.json --md-out blocked.md
node lib/merge-catalog-shards.js --raw actions-catalog.raw.json --shards shards-dir --out merged.json --dashboard dashboard.md
node lib/catalog-promotion-guard.js propose --catalog actions-catalog.raw.json --registry integration-library/registry.yaml --key example-action-key --proof proof.json --out registry.yaml
node lib/catalog-promotion-guard.js promote --catalog actions-catalog.raw.json --registry integration-library/registry.yaml --key example-action-key --proof proof.json --out registry.yaml
catalog-coverage-dashboard.js compares the raw catalog against the strict proof
registry and shows what's covered vs missing.catalog-promotion-guard.js propose|promote is how an action key moves into the
strict registry once you have proof it works - propose stages it, promote
commits it.merge-catalog-shards.js merges parallel discovery shard outputs (e.g. from a
Sandcastle-style parallel discovery run) into one catalog file.node clay-v2.js onboard-workspace --workspace {{workspace_id}} --workbooks wb_a,wb_b --limit 20 --update-library --out-dir onboard-output
Read-only against Clay (it writes local artifacts, never mutates the workspace), but
requires --update-library to also touch the local integration library files.
integration-library/registry.yaml ships with 40 Clay actions (plus per-action
templates in integration-library/templates/), each with observed input bindings,
status semantics, and a battle-test verdict. Observation evidence is intentionally
not included in the public build - validate anything you rely on with a <=10-row
sandbox run in your own workspace. If you ever hit ENOENT on
integration-library/registry.yaml, your checkout is incomplete - re-clone rather
than hand-creating the file.
node clay-v2.js integration-list --out integrations.json
node clay-v2.js integration-show example-action-key --out integration.json
node clay-v2.js integration-promotion-report --format markdown --out promotion-report.md
node clay-v2.js integration-validate-spec table.yaml --out validation.json
--confirm.action-template-generator.js and catalog-promotion-guard.js both work off
whatever catalog file you point them at - keep raw and normalized catalog files
separate so you don't feed a normalized file where a raw one is expected (check
each script's own usage string if unsure).generate-package-playbooks.js and package-rollup-report.js (per-package catalog
rollups) default to a repo-relative runs/2026-06-09/... path that simply won't
exist on a fresh clone - pass --catalog/--input explicitly rather than relying
on the default.