releases

Routine maintenance of indexed organizations — fetch all sources, regenerate overviews, verify data quality. Use when asked to "update", "refresh", or "maintain" one or more orgs, or when doing periodic sweeps of the registry.

Generate or refresh an org's AI overview from recent releases. Use when asked to "regenerate", "rewrite", or "refresh the overview" for one or more orgs, or as the regen step inside a maintaining-orgs run.

How to find, evaluate, and recommend the best ingestion method for changelog URLs — covers feed discovery, provider detection, GitHub API, markdown sources, and scraping fallback

Backfill a changelog source's full history locally in Claude Code via the backfill-source / backfill-sweep dynamic Workflows — preflight-gated, window-capped, budget-gated extraction written through the idempotent /batch upsert, with no remote extraction inference bill. Use when a source has a lot of history to pull in and running the remote update workflow per window would be too expensive. Local Claude Code only.

Put a challenge-blocked or unreliable scrape source on the external Firecrawl monitoring backend, or triage one that's already on it. Use when a page sits behind a bot challenge our Browser Rendering can't clear, when render keeps returning an empty shell, when enabling/disabling/tuning a monitor, or when a Firecrawl-backed source went quiet or ingested wrong content. Operator skill; requires admin API access and prod-only Firecrawl bindings.

How to create, delete, list, validate, and manage changelog sources — covers ignored/blocked URLs, duplicate detection, and the validation workflow

Group releases that cover the same underlying launch so readers see one story instead of three. Use when manually curating an org's coverage — reconciling a window of recent releases, fixing a bad automatic grouping, or linking coverage the deterministic clusterer can't see. Decide which item is the best entry point for the average reader and mark it canonical. Local Claude Code operator skill; ingest-time grouping is deterministic code, not this skill.

Generate the AI fields on releases — `title_generated`, `title_short`, `summary`, and `composition` — for one release, a window, or a backfill, via Claude Code sub-agents and direct SDK calls. Use when iterating on prompts, running provider/model experiments, or filling in a small-to-medium backfill outside the production ingest path or the Batches API. (Managed agents can spawn sub-agents now too; the dispatch scaffolding for this skill just isn't wired up on that side yet.)

Onboard or backfill a company's changelog locally in Claude Code by fetching pages and extracting releases with the agent itself (and parallel sub-agents), then writing through the batch-upsert endpoint — no remote fetch dispatch, no server-side extraction inference billing. Use when you want to onboard a source without triggering the remote update workflow's billed extraction, or recover from a remote fetch that burned tokens and wrote zero releases. Includes a mandatory robots.txt / Content-Signal opt-out gate.

Coordinate bulk playbook writing using parallel sub-agents — covers org discovery, prompt templates, model selection, batch dispatch, verification, and the parent-saves pattern for working around subagent permission limits. Currently wired up for Claude Code's Agent tool; the managed-agent fleet supports sub-agents too and the pattern can be adapted, but isn't wired up there yet.

Create or update a releases.json manifest so registries and agents can find where a company or project publishes its release notes. Use this whenever someone wants to "list our product on Releases", "add a releases.json", "get indexed on releases.sh", "here's our website, make us a manifest", declare a changelog / updates page / RSS feed / GitHub releases / App Store listing / CHANGELOG for discovery, publish a /.well-known/releases.json, or edit an existing one — even if they don't say "releases.json" by name. The usual input is just a company website or domain (no codebase required); the manifest is a small, schema-validated file the owner hosts on their domain or repo. This skill discovers the real publish locations, models products without over-fragmenting, writes the correctly-scoped file, validates it locally, and guides publishing it.

Decide whether an image or video found on a release page is editorial content (screenshots, demos, diagrams, product shots) or site chrome (avatars, logos, tracking pixels, decorative badges). Used during parsing to populate a release's media array.

How the Releases fetch and parse pipeline works — covers feed vs scrape adapters, incremental vs bulk parsing, dry-run testing, crawl mode, content hashing, and enrichment