Skip to main content
skill CX Codex

document-api-endpoint — Claude Code skill: what it does and setup

Documents and types Sentry API routes using drf-spectacular. It resolves type drift between declared OpenAPI schemas and live endpoint responses.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add getsentry/skills --skill document-api-endpoint -a claude-code
Open on GitHub

What document-api-endpoint does

The skill guides adding @extend_schema decorators to Sentry endpoint classes and methods. It reuses parameters from src/sentry/apidocs/parameters.py, assigns tags from OPENAPI_TAGS, and builds response shapes with TypedDict. For payload fields, T | None signals a required key with a null value, while NotRequired[T] handles optional keys like those triggered by an expand query parameter.

Migrating legacy paths is strictly all-or-nothing per path because drf-spectacular's APPEND_PATHS does not merge HTTP methods. Adding @extend_schema to one method causes all untransferred legacy methods on that route to disappear from the schema. Moving an endpoint to PUBLIC requires updating publish_status, assigning owner = ApiOwner.<TEAM>, and clearing the method from API_OWNERSHIP_ALLOWLIST_DONT_MODIFY.

Verification relies on make build-api-docs, pnpm run validate-api-examples, and targeted pytest runs in the virtual environment. To catch discrepancies like numbers formatted as strings, compare live curl output piped into jq against your types.

What it helps with

  • add @extend_schema decorators to a Sentry endpoint
  • fix type drift between runtime API responses and schemas
  • migrate legacy JSON specifications from api-docs/paths/
  • promote a private endpoint to PUBLIC with team ownership

Example prompts for the agent

  • Document this Sentry endpoint and write the response TypedDict
  • Fix type drift for this route against the live API response
  • Promote this endpoint to PUBLIC and remove it from the ownership allowlist

How to install document-api-endpoint

1Skills CLI (Vercel)

npx skills add getsentry/skills --skill document-api-endpoint -a claude-code

2Manually: put the files in place

~/.claude/skills/document-api-endpoint/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/document-api-endpoint/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about document-api-endpoint

How to install document-api-endpoint in Claude Code?

Run npx skills add getsentry/skills --skill document-api-endpoint -a claude-code in your terminal.

What does document-api-endpoint do?

It provides instructions for documenting Sentry endpoints with drf-spectacular. It covers response typing, decorator setup, parameter reuse, and schema validation.

Why must legacy api-docs paths be migrated all at once?

drf-spectacular's APPEND_PATHS does not merge HTTP methods on the same route. Once any method uses @extend_schema, all legacy methods on that path vanish unless migrated in the same commit.

Source text

SKILL.md · Apache-2.0
---
name: document-api-endpoint
description: Document and type a Sentry API endpoint. Write or fix @extend_schema decorators, specify response TypedDicts, type request parameters, correct type drift between the declared schema and the runtime response, and validate the generated spec. Use when asked to "document an endpoint", "add OpenAPI docs", "add/fix @extend_schema", "type an endpoint response", "fix the response type", "fix type drift", "reuse a response type", "split an overloaded endpoint", "specify the response schema", "add a TypedDict response", "migrate a legacy api-docs path", "fix a parameter type", or "make an endpoint public" / "promote an endpoint" (promotion is one section here).
---

# Document & Type a Sentry API Endpoint

Add or fix OpenAPI docs for a Sentry endpoint with drf-spectacular. Full reference is at https://develop.sentry.dev/backend/api/public/, the most useful section to you will be https://develop.sentry.dev/backend/api/public/#5-method-decorator. This skill captures the non-obvious lessons on top of it. Most of the work is making the declared schema match what the endpoint actually returns. Before documenting, identify which endpoint class serves the route and what it does; the MCP tool that calls it is usually the fastest way to confirm its behavior. Promoting a PRIVATE/EXPERIMENTAL endpoint to PUBLIC is one application (see below).

## Workflow

1. Class-level `@extend_schema(tags=[...])` — use the closest existing `OPENAPI_TAGS` entry.
2. Method-level `@extend_schema(operation_id=..., parameters=[...], responses={...}, examples=...)`.
3. Reuse `src/sentry/apidocs/parameters.py` and `examples/*.py`; ensure `owner = ApiOwner.<TEAM>` is set.
4. If a legacy `api-docs/paths/**/*.json` covers the path, remove it (see lesson 4).
5. Validate, then verify against the live endpoint (lesson 1).

## Lessons

### 1. Carefully compare what the code does vs declared types
Ideally, hit the live endpoint with a real token and diff the keys and types against your TypedDict. Serializers are sometimes inaccurate. Look out for counts coming back as floats instead of integers, IDs declared `int` emitted as strings, nested types declaring the wrong number of fields. Correct the declared type to match runtime.
```bash
curl -s -H "Authorization: Bearer $TOKEN" "https://us.sentry.io/api/0/<endpoint>" | jq 'keys'
```

### 2. Reuse the canonical response type
Match the codebase's `XxxResponseOptional(TypedDict, total=False)` mixin (main class declares required fields). Nullable-vs-absent: `T | None` = key always present, value may be null; `NotRequired[T]` = key only set under a condition (e.g. an `expand` query param). Reuse the existing canonical type instead of re-declaring a second or third copy in a `*_types.py`. If there's no clean canonical type to reuse (e.g. a payload proxied from another service like vroom/profiling), type it `dict[str, Any]` rather than inventing a new mirror, and confirm the shape from the owning service's repo, not just the serializer.

### 3. Infer the type. Avoid `cast` and `# type: ignore`
When a serializer returns a base type plus extra fields, refactor the producing code so the response type is inferred rather than forced.

### 4. Legacy doc migration is all-or-nothing per path
Delete the `api-docs/paths/**/*.json` file AND its `$ref` in `api-docs/openapi.json`. drf-spectacular's `APPEND_PATHS` does not merge HTTP methods, so once any method on a path uses `@extend_schema`, all *legacy* methods on that path vanish — migrate every method on the path in one commit.

## Promoting to PUBLIC

Do the workflow above, then on the concrete endpoint only (leave siblings PRIVATE):

- Bump `publish_status[<METHOD>]` → `PUBLIC` and set `owner = ApiOwner.<TEAM>`.
- Remove the method from `API_OWNERSHIP_ALLOWLIST_DONT_MODIFY` in the same change as the flip.
- If the endpoint is redundant or being renamed, delete or deprecate the old version in its own change first, then stack the publish on top.
- Note in the PR if scopes widen (`event:read` → `event:{admin,read,write}`) — that's drf-spectacular regenerating from `permission_classes`, documentation-only.

The change reaches the `@sentry/api` SDK / MCP only after `sentry-api-schema` regenerates downstream.

## Validate

```bash
make build-api-docs
pnpm run validate-api-examples
.venv/bin/pytest -q --reuse-db tests/apidocs/endpoints/<area>/test_<name>.py
.venv/bin/prek run -q --files <changed paths>
```

© getsentry · Apache-2.0 · getsentry/skills

Similar skills

Everything for Backend and APIs →

From the same set20

presentation-creator Creates interactive React and Vite slide decks with Sentry styling. Use it when a presentation needs real metrics, charts, and slide navigation. CX Codex +1 2,900installs django-perf-review This skill reviews Django code for N+1 queries, unbounded querysets and missing indexes. Useful before a release or when a page slows down on large tables. CX Codex +1 2,900installs django-access-review A skill that hunts IDOR and broken authorization in Django: it walks views, DRF viewsets and ORM queries and shows where checks are missing. CX Codex +1 2,700installs brand-guidelines Applies Sentry brand and copy guidelines to user-facing text, error states, and UI flows. Useful when writing UI labels, empty states, or docs. CX Codex +1 2,500installs skill-writer Designs, synthesizes, and refines agent skills following the Agent Skills specification. Helpful when turning raw workflows or prompt templates into structured skills. CX Codex +1 2,500installs prompt-optimizer A skill that rewrites system and agent prompts against a set of eval cases instead of gut feeling. Useful when a prompt exists but behaves inconsistently. CX Codex +1 2,300installs sred-project-organizer Organizes annual engineering projects into SRED-compliant documentation and creates structured Notion pages. CX Codex +1 2,100installs typing-exclusion-worker Removes Python modules from mypy exclusion lists in pyproject.toml, fixes typing errors, and runs scoped validation. Built for parallel typing debt workers. CX Codex +1 1,600installs

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor