Inputs & Outputs

Reference documentation for all Warden GitHub Action inputs and outputs.

Inputs

All inputs are configured in your workflow file under with::

- uses: getsentry/warden@v1
  with:
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}
    fail-on: high

anthropic-api-key

string

Anthropic API key (sk-ant-...) or OAuth token (sk-ant-oat-...).Can also be set via environment variables:

WARDEN_ANTHROPIC_API_KEY
ANTHROPIC_API_KEY
CLAUDE_CODE_OAUTH_TOKEN (for OAuth)

Required: One authentication method must be provided.

# Via input
- uses: getsentry/warden@v1
  with:
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}

# Via environment variable
- uses: getsentry/warden@v1
  env:
    ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

github-token

string

default:"${{ github.token }}"

GitHub token for API access. Defaults to the automatic GITHUB_TOKEN.Use a GitHub App token for better review thread management:

- uses: actions/create-github-app-token@v1
  id: app-token
  with:
    app-id: ${{ secrets.WARDEN_APP_ID }}
    private-key: ${{ secrets.WARDEN_PRIVATE_KEY }}

- uses: getsentry/warden@v1
  with:
    github-token: ${{ steps.app-token.outputs.token }}

config-path

string

default:"warden.toml"

Path to warden.toml configuration file, relative to repository root.

- uses: getsentry/warden@v1
  with:
    config-path: .github/warden.toml

fail-on

string

default:"high"

Minimum severity level to fail the action.Options: off, critical, high, medium, low, infoWhen findings at or above this severity are found:

Action fails if fail-check is true
Review is REQUEST_CHANGES if request-changes is true

Use off to never fail (informational mode).

- uses: getsentry/warden@v1
  with:
    fail-on: critical  # Only fail on critical issues

report-on

string

default:"medium"

Minimum severity level to show annotations in code review.Options: off, critical, high, medium, low, infoFindings below this threshold are counted but not posted as comments. Use off to disable all review comments.

- uses: getsentry/warden@v1
  with:
    report-on: low  # Show all findings in review

max-findings

number

default:"50"

Maximum number of findings to report. Use 0 for unlimited.Higher severity findings are prioritized when limit is reached.

- uses: getsentry/warden@v1
  with:
    max-findings: 100  # Allow up to 100 findings

request-changes

boolean

default:"false"

Whether to use REQUEST_CHANGES review event when findings exceed fail-on threshold.When true:

Creates a blocking review that requires dismissal
Automatically dismissed when all issues are resolved
Requires GitHub token with reviewer permissions

- uses: getsentry/warden@v1
  with:
    fail-on: high
    request-changes: true  # Block PR merge

fail-check

boolean

default:"false"

Whether to fail the check run when findings exceed fail-on threshold.When true:

Check run status is “failure”
Blocks PR merge if check is required

When false:

Check run always passes
Findings shown for information only

- uses: getsentry/warden@v1
  with:
    fail-check: true  # Fail check on high+ severity

parallel

number

default:"5"

Maximum number of concurrent trigger executions.Higher values:

Faster analysis for large changes
Higher API usage and memory

Lower values:

Slower but more conservative
Better for rate limit management

- uses: getsentry/warden@v1
  with:
    parallel: 10  # Analyze 10 files concurrently

Outputs

Access outputs in subsequent workflow steps:

- uses: getsentry/warden@v1
  id: warden
  with:
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}

- name: Check results
  run: |
    echo "Found ${{ steps.warden.outputs.findings-count }} issues"
    echo "High severity: ${{ steps.warden.outputs.high-count }}"

findings-count

number

Total number of findings across all triggers and severity levels.Includes all findings regardless of report-on threshold.

- name: Post results to Slack
  if: steps.warden.outputs.findings-count > 0
  run: |
    curl -X POST ${{ secrets.SLACK_WEBHOOK }} \
      -d '{"text": "Warden found ${{ steps.warden.outputs.findings-count }} issues"}'

high-count

number

Number of high severity findings.Useful for tracking critical issues separately.

- name: Fail on critical issues
  if: steps.warden.outputs.high-count > 5
  run: exit 1

summary

string

Human-readable summary of the analysis.Example: "Found 3 issues: 1 high, 2 medium"

- name: Comment summary
  uses: actions/github-script@v7
  with:
    script: |
      github.rest.issues.createComment({
        issue_number: context.issue.number,
        owner: context.repo.owner,
        repo: context.repo.repo,
        body: 'Warden: ${{ steps.warden.outputs.summary }}'
      })

findings-file

string

Path to structured JSON findings file.Always written, even when no findings are found. Use for:

Uploading to cloud storage (GCS, S3)
Custom processing pipelines
Archiving results

Format: See Findings Schema

- name: Upload findings
  uses: actions/upload-artifact@v4
  with:
    name: warden-findings
    path: ${{ steps.warden.outputs.findings-file }}

- name: Upload to GCS
  run: |
    gsutil cp ${{ steps.warden.outputs.findings-file }} \
      gs://my-bucket/warden/${{ github.sha }}.json

Findings Schema

The JSON file referenced by findings-file has this structure:

{
  "event": {
    "type": "pull_request",
    "action": "opened",
    "repository": {
      "owner": "getsentry",
      "name": "warden",
      "fullName": "getsentry/warden"
    },
    "pullRequest": {
      "number": 123,
      "baseBranch": "main",
      "headBranch": "feature/fix",
      "baseSha": "abc123",
      "headSha": "def456"
    }
  },
  "reports": [
    {
      "skill": "security-audit",
      "findings": [
        {
          "id": "sec-001",
          "severity": "high",
          "confidence": "high",
          "title": "SQL Injection Risk",
          "message": "User input is directly interpolated into SQL query",
          "path": "src/db/users.ts",
          "line": 42,
          "suggestion": "Use parameterized queries instead"
        }
      ],
      "duration": 1234,
      "usage": {
        "inputTokens": 5000,
        "outputTokens": 1000,
        "costUSD": 0.05
      }
    }
  ],
  "summary": {
    "totalFindings": 5,
    "bySeverity": {
      "critical": 0,
      "high": 1,
      "medium": 3,
      "low": 1,
      "info": 0
    },
    "totalCostUSD": 0.15,
    "totalDurationMs": 3500
  }
}

Usage Examples

Conditional Workflows

Run different checks based on findings:

jobs:
  warden:
    runs-on: ubuntu-latest
    outputs:
      has-findings: ${{ steps.warden.outputs.findings-count > 0 }}
      high-count: ${{ steps.warden.outputs.high-count }}
    steps:
      - uses: actions/checkout@v4
      - uses: getsentry/warden@v1
        id: warden
        with:
          anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}
  
  notify:
    needs: warden
    if: needs.warden.outputs.has-findings == 'true'
    runs-on: ubuntu-latest
    steps:
      - name: Notify team
        run: echo "Found issues, notifying team..."
  
  security-review:
    needs: warden
    if: needs.warden.outputs.high-count > 0
    runs-on: ubuntu-latest
    steps:
      - name: Request security review
        run: echo "High severity issues found, requesting security review..."

Upload Findings to Cloud Storage

- uses: getsentry/warden@v1
  id: warden
  with:
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}

- name: Upload to S3
  if: always()
  env:
    AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
    AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
  run: |
    aws s3 cp ${{ steps.warden.outputs.findings-file }} \
      s3://my-warden-reports/${{ github.repository }}/${{ github.sha }}.json

Custom Summary Comment

- uses: getsentry/warden@v1
  id: warden
  with:
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}

- name: Post custom summary
  if: always()
  uses: actions/github-script@v7
  with:
    script: |
      const summary = `## Warden Analysis
      
      ${{ steps.warden.outputs.summary }}
      
      - Total findings: ${{ steps.warden.outputs.findings-count }}
      - High severity: ${{ steps.warden.outputs.high-count }}
      
      [View detailed findings](${{ steps.warden.outputs.findings-file }})`;
      
      github.rest.issues.createComment({
        issue_number: context.issue.number,
        owner: context.repo.owner,
        repo: context.repo.repo,
        body: summary
      });

Matrix Testing

Run Warden with different configurations:

jobs:
  warden:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        config:
          - { fail-on: 'critical', name: 'Critical Only' }
          - { fail-on: 'high', name: 'High+' }
          - { fail-on: 'medium', name: 'Medium+' }
    steps:
      - uses: actions/checkout@v4
      - uses: getsentry/warden@v1
        with:
          anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}
          fail-on: ${{ matrix.config.fail-on }}
          fail-check: false  # Don't block, just report

Get Started

Core Concepts

CLI Usage

GitHub Action

Configuration

Skills

Advanced

Inputs & Outputs

Inputs

anthropic-api-key

github-token

config-path

fail-on

report-on

max-findings

request-changes

fail-check

parallel

Outputs

findings-count

high-count

summary

findings-file

Findings Schema

Usage Examples

Conditional Workflows

Upload Findings to Cloud Storage

Custom Summary Comment

Matrix Testing

Next Steps

Setup

Configuration

Build docs developers (and LLMs) love

Get Started

Core Concepts

CLI Usage

GitHub Action

Configuration

Skills

Advanced

​Inputs

​anthropic-api-key

​github-token

​config-path

​fail-on

​report-on

​max-findings

​request-changes

​fail-check

​parallel

​Outputs

​findings-count

​high-count

​summary

​findings-file

​Findings Schema

​Usage Examples

​Conditional Workflows

​Upload Findings to Cloud Storage

​Custom Summary Comment

​Matrix Testing

​Next Steps

Setup

Configuration

Build docs developers (and LLMs) love

Inputs

anthropic-api-key

github-token

config-path

fail-on

report-on

max-findings

request-changes

fail-check

parallel

Outputs

findings-count

high-count

summary

findings-file

Findings Schema

Usage Examples

Conditional Workflows

Upload Findings to Cloud Storage

Custom Summary Comment

Matrix Testing

Next Steps