# IonWarp Docs

> Beyond AI Code Review. Machine-readable docs for IonWarp — 53 pages. Every capability has a human door and an agent door: same gates, same objects, same timeline.

## Introduction

- [What IonWarp is](https://ionwarp.com/docs): IonWarp is a repo-quality agent: it reviews every pull request, posts what it found on the PR, and scores your repository's readiness.
- [Build with AI](https://ionwarp.com/docs/start/build-with-ai): Describe the repository once; IonWarp seeds the review agents, scorecards, and PR sync.
- [Connect Your Agent](https://ionwarp.com/docs/start/connect-agent): Point a coding agent, a business agent, or any MCP client at IonWarp — one server URL, one token, three ways to wire it.

## Get Started

- [Connect GitHub](https://ionwarp.com/docs/get-started/connect-github): Install the app, pick repositories, and read the honest empty scorecard.
- [Add Agents](https://ionwarp.com/docs/get-started/add-agents): Curate the reviewer roster your reviews run with.
- [Your First Review](https://ionwarp.com/docs/get-started/first-review): The review lifecycle, and where its output lands.
- [Your First Scorecard](https://ionwarp.com/docs/get-started/first-scorecard): Checks, levels, and how the score ratchets.
- [Add IonWarp to Agents.md](https://ionwarp.com/docs/get-started/agents-md): One PR that wires your coding agents into IonWarp reviews — Autofix + Session Analytics.

## API

- [API Quickstart](https://ionwarp.com/docs/api/quickstart): Authenticate, name your project, and make your first call against the IonWarp API.
- [Repositories & PRs](https://ionwarp.com/docs/api/repos): Connect a repository through the GitHub App, watch its pull requests, and read the PR projection every review is keyed to.
- [Reviews](https://ionwarp.com/docs/api/reviews): One review pass over a pull request: launch it, read the per-reviewer telemetry, and see the outcome stamped on the PR.
- [Reviewers & Models](https://ionwarp.com/docs/api/reviewers): The reviewer catalog, explicit lineups with per-reviewer model overrides, the three-level model resolution order, and how a trial winner becomes a project pin.
- [Scorecards](https://ionwarp.com/docs/api/scorecards): Repeatable quality rubrics scored against the connected repo — run a scan, read levels and checks, and track the score by model over time.
- [Plans](https://ionwarp.com/docs/api/plans): Proposed work — discovered from the repo or written by IonWarp — scored, reviewed and approved before anything is written back, plus the learnings that inform it.
- [Checklists](https://ionwarp.com/docs/api/checklists): Read a project's checklists — onboarding and the daily Beast Mode operations bar that grades every review lens on HSF-E — with per-day history.
- [Control Plane](https://ionwarp.com/docs/api/control-plane): Projects, run modes, the approval brake, and the spend cap — the operator controls that decide whether a run may start, and the named refusals when one does not.
- [Debugging a Run](https://ionwarp.com/docs/api/debugging-a-run): Get a run id, read the status block, tell success from a visible failure, and name the failure class — without guessing from silence.
- [Platform](https://ionwarp.com/docs/api/platform): An annotated index of the platform contract every IonWarp surface shares — tasks, data, timeline, usage, integrations, skills, and the cross-cutting reference pages.
- [MCP Tools](https://ionwarp.com/docs/reference/mcp-tools): Every mcp__dash__* tool available to IonWarp, with arguments and a minimal call.

## Skills

- [Skills](https://ionwarp.com/docs/skills): The ionwarp skill catalog.

## Reference

- [API Resources](https://ionwarp.com/docs/collections): Every REST resource ionwarp serves, with its operations.
- [Integrations](https://ionwarp.com/docs/integrations): Every provider this product connects to.
- [llms.txt](https://ionwarp.com/docs/reference/llms-txt): The machine index of these docs — every page, as plain text, for an agent to read.

## Other pages

Served and linked from the pages above, but not in the sidebar.

- [Agent Onboarding](https://ionwarp.com/docs/agent-onboarding): The paste-once bootstrap — everything an arbitrary agent needs to start driving this agent, in one fetch.
- [Checks](https://ionwarp.com/docs/collections/checks): The individual checks a scorecard scores.
- [Dashboards](https://ionwarp.com/docs/collections/dashboards): Metric boards built from your repo's review history.
- [Findings](https://ionwarp.com/docs/collections/findings): Structured review findings — the durable ledger the website renders.
- [Integrations](https://ionwarp.com/docs/collections/integrations): Provider credentials attached to a project.
- [Learnings](https://ionwarp.com/docs/collections/learnings): What IonWarp learned about this repo, with an expiry.
- [Metrics](https://ionwarp.com/docs/collections/metrics): Stored metric values — one row per (owner, key), newest write wins.
- [Plans](https://ionwarp.com/docs/collections/plans): Proposed work, scored and approved before anything is written.
- [Preferences](https://ionwarp.com/docs/collections/preferences): Per-project dashboard state (notification read/cleared markers).
- [Pull Requests](https://ionwarp.com/docs/collections/pull-requests): Pull requests IonWarp is watching, and where each one got to in review.
- [Quick Actions](https://ionwarp.com/docs/collections/quick-actions): One-click next steps the scorecard suggests.
- [Reviewers](https://ionwarp.com/docs/collections/reviewers): The reviewers available to run on your repo.
- [Pull Requests](https://ionwarp.com/docs/collections/reviews): A single review pass over a pull request.
- [Scorecard Levels](https://ionwarp.com/docs/collections/scorecard-levels): The ladder a repo climbs, level by level.
- [Scorecards](https://ionwarp.com/docs/collections/scorecards): Your repo's readiness score and how it was earned.
- [Skills](https://ionwarp.com/docs/collections/skills): Skills this agent can run, and where their executed bytes come from.
- [Review Swarms](https://ionwarp.com/docs/collections/swarms): Multi-agent runs coordinated toward one goal.
- [Timeline Events](https://ionwarp.com/docs/collections/timeline-events): The append-only record of what happened to an object.
- [Data](https://ionwarp.com/docs/core/data): Read and write objects in a project's Data collections with a PAT — data_set, data_get, data_query, data_batch. Every field name is snake_case.
- [Integrations](https://ionwarp.com/docs/core/integrations): Provider credentials, the credential-blind call proxy, the capability catalog, and risk classes — discover the contract, never hardcode a provider.
- [Tasks](https://ionwarp.com/docs/core/tasks): Create, run, and read tasks — every unit of agent work runs as a task with one status read and one transcript read.
- [Timeline](https://ionwarp.com/docs/core/timeline): Append and read human-readable events on any object — events, provenance, and timestamp touches.
- [Usage](https://ionwarp.com/docs/core/usage): Read a project's credit spend and the per-run itemized cost — usage_summary and the usage_line_items block on a run's status.
- [Authentication](https://ionwarp.com/docs/reference/auth): Per-route authentication modes and required scopes.
- [Conventions](https://ionwarp.com/docs/reference/conventions): Cross-cutting API conventions — id prefixes, snake_case fields, epoch-millisecond time, and credit-denominated money.
- [Error Codes](https://ionwarp.com/docs/reference/errors): The canonical REST error codes and their default HTTP status.
- [Pagination](https://ionwarp.com/docs/reference/pagination): List endpoints return a bounded page — pass limit and offset, and read has_more to decide whether to ask for another page.
- [review-study](https://ionwarp.com/docs/skills/review-study): Daily quality grading for IonWarp's PR reviews. The head-to-head half (§0) is the point of the skill: pick ONE real PR where at least two review bots commented, read what each bot ACTUALLY WROTE, and grade them head-to-head against our own rubric — judging false positives and MISSED bugs, which no counter can see. §1-§5 add the fleet reliability strip (n, err%, latency, cost) from reviewers_telemetry. Never uses IonWarp's own self-grade as a quality verdict — that number counts findings, and volume is not quality. Read-only — never launches reviews, never writes pins, never posts to GitHub.
- [reviewer-model-lab](https://ionwarp.com/docs/skills/reviewer-model-lab): The IonWarp reviewer model lab. Per run: at most TWO trials against the newest review-study snapshot — trial 1 chases an open lead (or grading's top recommendation), trial 2 is one web-researched frontier candidate confirmed on OpenRouter. REPORT ONLY: it measures and recommends, and NEVER changes a pin — a model switch is an operator decision, proposed in the digest with its evidence.
- [scorecard-model-lab](https://ionwarp.com/docs/skills/scorecard-model-lab): The IonWarp scorecard lab — one run across the configured projects: ≤2 refresh slots (never-run > restore > stale > top-up) on the resolved pin, then ≤1 model trial against a ≤7-day committed baseline. A scan counts only when its projection COMMITTED. REPORT ONLY — it proposes pin changes with their evidence and never writes one; a paused project is never unblocked.

## Installable skills

Executable operator procedures, published with a content hash. To use one:

1. Fetch `https://ionwarp.com/docs/skills/index.json`.
2. For each skill you want, compare its `version` to your cached copy's.
   `version` is the sha256 of `SKILL.md` and equals the hash the platform
   records as that skill's `executed_hash` — identical version, identical
   bytes, no refetch needed.
3. On a change, fetch `skill_url` (the executable `SKILL.md`), or download
   `zip_url` / each `files[].url` for a multi-file skill. Record the new
   `version`.
4. `SKILL.md` is the executable procedure; `references/*` are its contracts
   and `scripts/*` are read-only helpers. Follow the body — it names the
   live doc pages to read rather than restating them.

- [review-study](https://ionwarp.com/docs/skills/review-study): Daily quality grading for IonWarp's PR reviews. The head-to-head half (§0) is the point of the skill: pick ONE real PR where at least two review bots commented, read what each bot ACTUALLY WROTE, and grade them head-to-head against our own rubric — judging false positives and MISSED bugs, which no counter can see. §1-§5 add the fleet reliability strip (n, err%, latency, cost) from reviewers_telemetry. Never uses IonWarp's own self-grade as a quality verdict — that number counts findings, and volume is not quality. Read-only — never launches reviews, never writes pins, never posts to GitHub. (version `ab0999492611…`)
- [reviewer-model-lab](https://ionwarp.com/docs/skills/reviewer-model-lab): The IonWarp reviewer model lab. Per run: at most TWO trials against the newest review-study snapshot — trial 1 chases an open lead (or grading's top recommendation), trial 2 is one web-researched frontier candidate confirmed on OpenRouter. REPORT ONLY: it measures and recommends, and NEVER changes a pin — a model switch is an operator decision, proposed in the digest with its evidence. (version `a1fb433f3d0d…`)
- [scorecard-model-lab](https://ionwarp.com/docs/skills/scorecard-model-lab): The IonWarp scorecard lab — one run across the configured projects: ≤2 refresh slots (never-run > restore > stale > top-up) on the resolved pin, then ≤1 model trial against a ≤7-day committed baseline. A scan counts only when its projection COMMITTED. REPORT ONLY — it proposes pin changes with their evidence and never writes one; a paused project is never unblocked. (version `ad761a3595b3…`)

## Operational notes

- Raw markdown for any page above: `https://ionwarp.com/docs/raw/<slug>.md`
- Machine index of every page: `https://ionwarp.com/docs/index.json`
- Machine index of every skill: `https://ionwarp.com/docs/skills.json` (also at `https://ionwarp.com/.well-known/skills/index.json` on the docs host)
- The whole corpus as one document: `https://ionwarp.com/docs/llms-full.txt`
- This file: `https://ionwarp.com/docs/llms.txt`
- API base: `https://ionwarp.com/api/v1`
- Auth: `Authorization: Bearer $DASH_API_KEY` (a `pat_…` workspace token), and name the project with the `Project-ID` header.
- Bootstrap an agent in one paste: Read and follow https://ionwarp.com/docs/raw/agent-onboarding.md