The pre-research scaffold. Decomposes a question into sub-questions so the synthesis step has a structure to fill.

The first step in the decision pipeline. Enumerates the named camps on a contested topic so downstream skills can survey sources per camp and compare hypotheses across them.

Use this skill to walk the adopter through the four classes of post-install change (substitution, .upstream companion merges, discovery + restructuring, within-layout consolidation). Triggers after installing a pack (the install->adapt chain nudges via session-start hook) or any time `<repo>/.adapt-install-marker.toml` / `~/.agentbundle/.adapt-install-marker.toml` is on disk. Walks both scopes' state files for Tier-2 detection; class-1 substitution shells out to `agentbundle adapt`; classes 2-4 write files directly under the per-scope path-jail.

Use when generating an OpenAPI 3.1 API contract from requirements, user stories, or domain models. Applies a pluggable API standard (Zalando by default) as hard constraints to produce complete, validated YAML specs ready for code gen, test gen, mocks, and SDKs. Activate for tasks involving API design, REST contract authoring, or OpenAPI spec creation.

Use to survey a whole external repo or catalogue (local path or URL) for ingestion candidates and produce a reviewable RFC of per-candidate verdicts (assimilate, reject, or needs-new-pack), resumable across sessions and git worktrees via a ledger. Triggers on "survey this repo for skills we can adopt", "inventory this catalogue", "what can we assimilate from <repo>", "re-sync from <upstream>". Do NOT use for a single known unit (use assimilate-primitive) or to scaffold a proposed pack (use propose-catalogue-pack).

Use this skill when the user has unstructured external input (an email thread, a prose description, a Linear Issue, a stakeholder message) and needs to produce a DoR-compliant product brief and queue it in workspace.toml. Triggers on "author a brief", "write a brief from this email", "create a brief from this Linear issue", "intake this brief", "turn this into a brief". Do NOT use to decompose an existing brief into specs (use receive-brief) or to author a single feature from scratch (use new-spec).

Create, revise, retrofit, audit, or verify product documentation — pack READMEs, journeys, tutorials, how-to guides, reference pages, and explanations. Use when asked to write, improve, restructure, audit, or verify user-facing documentation, fix a pack README, create a guide for a feature, update a journey page, or check whether docs match shipped behavior. Infers the mode from the request. Do NOT use for feature specifications (use new-spec), cross-cutting proposals (use new-rfc), decisions (use new-adr), product or market strategy, frontend implementation alone, internal maintainer runbooks without user-facing concern, or arbitrary prose editing with no documentation purpose.

A credentialed-CLI fixture skill that satisfies every lint check; lint should report zero findings.

Passive stop-signal for a research project — reads the synthesis matrix and memos by eye and reports whether the corpus has stopped changing the structure (theoretical saturation), plus a recommendation. Triggers on project-lifecycle phrasing — "is this project saturated", "should I keep gathering", "check the stop signal" — inside an existing project folder. It NEVER advances phase and computes no counter, score, or metric; the saturation judgment is qualitative and the human decides. It MAY optionally write a verdict_status string into overview.md (the single permitted light state write) — nothing more. Prompt-only by construction.

Orient to the current desk-research project at a glance — reads overview.md and surfaces phase, working hypothesis, stop-signal verdict, and what to do next. Triggers on 'where are we on the X research', 'status of the Y investigation', 'resume the Z project', 'what phase is the research in', 'where did we leave off on the research', or any return-to-a-named-research-project phrasing. Read-only: never advances phase, never modifies any file.

Use when designing a documentation surface — a docs site, a help centre, an API reference, or a technical guide set. Decides what type of content belongs where, how navigation scales with content volume, and what the first-value-moment is for each content type. Triggers on 'design the docs site', 'structure the help centre', 'what goes on the docs landing page', 'how should we navigate the API reference', 'TTFV for this tutorial'. Produces IA and navigation specifications for documentation surfaces. Do NOT use for marketing surfaces (use conversion-design) or informational editorial pages (use informational-design). Surface genre: documentation.

Credentialed fixture whose script opens `.agentbundle/credentials.env` directly without the opt-out marker; a dotfile-read finding is expected.

Credentialed primitive that legitimately reads the dotfile, with the opt-out marker on the same line; lint must NOT flag it.

Convert documents and images to Markdown for AI context layers. Documents (PDF, DOCX, XLSX, PPTX, HTML, EPUB, CSV/TSV, OpenDocument, .eml) go through `scripts/convert.py`, which extracts at a no-ML Tier-0 floor (pure-Python / stdlib parsers) and falls through to Docling (Tier 2) only for .xls and images; images go through a two-pass sliding-window vision pipeline whose tiling and reconciliation are deterministic (`scripts/split_image.py` and `scripts/reconcile.py`). Every output carries a versioned frontmatter contract (provenance + a quality/confidence signal). The agent's job is the per-tile vision read; tile dedup and ordering are handled by the script.

Use this skill when the user asks for DORA / Flow Framework metrics over a Jira scope -- "what's our cycle time this quarter for PROJ", "give me throughput and WIP for the Foo team", "compare flow efficiency before/after the AI-pairing rollout via a cohort split", "rollup flow metrics across program 42". Computes cycle time, lead time, throughput, WIP, flow load, rework rate, flow efficiency, flow distribution, and defect ratio from Jira changelogs (optionally joined with Jira Align for program / portfolio scope). Read-only -- never transitions, comments, creates, updates, or deletes Jira data. Do NOT use for live-dashboard streaming, deployment-event metrics (Change Failure Rate / MTTR proper), or anything requiring a tracker that isn't Jira.

**Use this when:** your skill calls an external authenticated service (API token, vendor CLI, or corporate SSO) and the credential must never pass through the model. **Prerequisites:** target service's auth shape identified, namespace name chosen, and the `credential-brokers` pack installed if you'r

Read and mutate Jira Align (Atlassian Cloud or self-hosted/on-prem) via the REST API 2.0. Supports fetching individual records (epics, features, stories, capabilities, themes, portfolios, programs, teams, users, etc.), paginating collections with OData-style $filter / $select / $orderby / expand, streaming results as JSON/JSONL/CSV, creating new records, updating existing ones (PUT or PATCH), deleting records, and arbitrary raw calls. Use when the user wants to read, search, export, create, or update Jira Align data.

Use this skill when the user points at a Jira defect and asks you to handle it end-to-end -- "fix PROJ-123", "work this bug ticket", "diagnose and ship a fix for this defect". The skill pulls the ticket via the `jira` skill, hands the actual fix to the `bug-fix` skill (reproduction-first, root-cause, minimum-diff, regression-test), opens a PR whose body links back to Jira, and comments + transitions the ticket. By default it stops at PR-opened; runs a dev-deploy step only if the consumer repo provides one. Do NOT use for stories, tasks, or feature work -- those go through `new-spec`.

Use when designing a marketplace surface — a catalogue, a listing grid, a product detail page, or a buying/transaction flow that connects buyers and sellers or producers and consumers. Triggers on 'design the marketplace', 'structure the listing page', 'how should the catalogue work', 'design the search and filter', 'buyer journey on the marketplace', 'product card design'. Produces IA specifications for catalogue, filter, comparison, and transaction bridge surfaces. Do NOT use for single-product marketing surfaces (use conversion-design) or workspace tool surfaces (use workspace-design). Surface genre: marketplace.

Use this skill when the user asks to create, write, draft, or open a new ADR (architecture decision record). Triggers on phrases like "new ADR", "write an ADR for...", "record this decision", "let's ADR this". Do NOT use for RFCs (use `new-rfc`) or feature specs (use `new-spec`).