Continuous Integration

Tracer integrates seamlessly with CI/CD pipelines to automate issue creation, track test failures, and coordinate agent work based on build results.

CI/CD Benefits

Using Tracer in continuous integration provides:

Automated issue creation from test failures
Build status tracking through comments
Agent coordination based on CI results
Audit trail of automated changes

All Tracer commands support --json output for easy parsing in CI/CD scripts.

Basic CI Integration

Minimal integration that creates issues from failures:

# .github/workflows/ci.yml
name: CI with Tracer

on: [push, pull_request]

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2
      
      - name: Install Tracer
        run: cargo install tracer-cli
      
      - name: Run tests
        id: tests
        run: |
          export TRACE_ACTOR="ci-bot"
          cargo test || echo "tests_failed=true" >> $GITHUB_OUTPUT
      
      - name: File issues for failures
        if: steps.tests.outputs.tests_failed == 'true'
        run: |
          export TRACE_ACTOR="ci-bot"
          tracer create "Build #${{ github.run_number }} failed" -t bug -p 0

JSON Output Parsing

Use jq to parse Tracer JSON output in CI scripts:

# Count ready issues
READY_COUNT=$(tracer ready --json | jq '. | length')
echo "Ready issues: $READY_COUNT"

Test Failure Tracking

Automatically create issues for test failures:

#!/bin/bash

export TRACE_ACTOR="ci-bot"
BUILD_NUMBER=${GITHUB_RUN_NUMBER:-"local"}

# Run tests and capture output
if ! cargo test 2>&1 | tee test_output.txt; then
  echo "Tests failed - processing failures"
  
  # Parse failed tests
  FAILED_TESTS=$(grep "FAILED" test_output.txt | awk '{print $2}')
  
  for test in $FAILED_TESTS; do
    ISSUE_TITLE="Test failure: $test"
    
    # Check if issue already exists
    EXISTING=$(tracer list --json | jq -r ".[] | select(.title == \"$ISSUE_TITLE\" and .status != \"closed\") | .id" | head -n 1)
    
    if [ -z "$EXISTING" ]; then
      # Create new issue
      NEW_ISSUE=$(tracer create "$ISSUE_TITLE" -t bug -p 0 --json | jq -r '.id')
      tracer comment $NEW_ISSUE "Failed in build #$BUILD_NUMBER"
      tracer comment $NEW_ISSUE "See: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
      echo "Created issue $NEW_ISSUE for $test"
    else
      # Update existing issue
      tracer comment $EXISTING "Still failing in build #$BUILD_NUMBER"
      echo "Updated $EXISTING - $test still failing"
    fi
  done
  
  exit 1
fi

echo "All tests passed - closing any test failure issues"

# Close resolved test failure issues
OPEN_FAILURES=$(tracer list --json | jq -r '.[] | select(.title | startswith("Test failure:") and .status != "closed") | .id')

for issue in $OPEN_FAILURES; then
  tracer close $issue --reason "Tests passing in build #$BUILD_NUMBER"
  echo "Closed $issue - tests now passing"
done

Build Status Comments

Add build status updates to relevant issues:

#!/bin/bash

export TRACE_ACTOR="ci-bot"
BUILD_URL="${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"

# Find issues mentioned in commits
ISSUE_IDS=$(git log -1 --pretty=%B | grep -oE '(bd-[0-9]+)' | sort -u)

for issue in $ISSUE_IDS; do
  # Verify issue exists
  if tracer show $issue --json > /dev/null 2>&1; then
    if [ "$BUILD_STATUS" = "success" ]; then
      tracer comment $issue "✓ Build passed: $BUILD_URL"
    else
      tracer comment $issue "✗ Build failed: $BUILD_URL"
    fi
  fi
done

Agent Task Assignment from CI

Assign work to agents based on CI results:

Detect issues in build

# CI detects linting errors
if ! cargo clippy -- -D warnings; then
  ISSUE=$(tracer create "Fix clippy warnings" -t chore -p 1 --json | jq -r '.id')
  tracer comment $ISSUE "$(cargo clippy 2>&1)"
fi

Assign to available agent

# Assign to agent with least work
for agent in claude-1 cursor-2 gpt-4; do
  COUNT=$(tracer list --assignee $agent --status in_progress --json | jq '. | length')
  echo "$agent: $COUNT issues"
done | sort -k2 -n | head -n 1 | awk '{print $1}' > /tmp/agent

AGENT=$(cat /tmp/agent)
tracer update $ISSUE --assignee $AGENT
tracer comment $ISSUE "Auto-assigned to $AGENT"

Notify agent

# Agent picks up work in next session
export TRACE_ACTOR=$AGENT
tracer list --assignee $TRACE_ACTOR

Dependency Updates

Track dependency update issues:

#!/bin/bash

export TRACE_ACTOR="dependabot"

# Check for outdated dependencies
OUTDATED=$(cargo outdated --format json | jq -r '.dependencies[] | select(.latest != .project) | .name')

for dep in $OUTDATED; do
  ISSUE_TITLE="Update dependency: $dep"
  
  # Check if issue exists
  EXISTING=$(tracer list --json | jq -r ".[] | select(.title == \"$ISSUE_TITLE\" and .status != \"closed\") | .id" | head -n 1)
  
  if [ -z "$EXISTING" ]; then
    CURRENT=$(cargo outdated --format json | jq -r ".dependencies[] | select(.name == \"$dep\") | .project")
    LATEST=$(cargo outdated --format json | jq -r ".dependencies[] | select(.name == \"$dep\") | .latest")
    
    NEW_ISSUE=$(tracer create "$ISSUE_TITLE" -t chore -p 2 --json | jq -r '.id')
    tracer comment $NEW_ISSUE "Current: $CURRENT, Latest: $LATEST"
    echo "Created $NEW_ISSUE for $dep update"
  fi
done

Performance Regression Detection

Create issues for performance regressions:

#!/bin/bash

export TRACE_ACTOR="perf-bot"

# Run benchmarks
cargo bench --bench main -- --save-baseline current

# Compare with previous baseline
if cargo bench --bench main -- --baseline previous 2>&1 | grep -q "regressed"; then
  REGRESSIONS=$(cargo bench --bench main -- --baseline previous 2>&1 | grep "regressed" | awk '{print $1}')
  
  for bench in $REGRESSIONS; do
    ISSUE_TITLE="Performance regression: $bench"
    
    EXISTING=$(tracer list --json | jq -r ".[] | select(.title == \"$ISSUE_TITLE\" and .status != \"closed\") | .id" | head -n 1)
    
    if [ -z "$EXISTING" ]; then
      NEW_ISSUE=$(tracer create "$ISSUE_TITLE" -t bug -p 1 --json | jq -r '.id')
      tracer comment $NEW_ISSUE "Detected in build #${BUILD_NUMBER}"
      echo "Created $NEW_ISSUE for regression in $bench"
    fi
  done
fi

Complete CI/CD Workflow

Full GitHub Actions workflow:

# .github/workflows/tracer-ci.yml
name: Tracer CI/CD Integration

on:
  push:
    branches: [main]
  pull_request:
    branches: [main]

env:
  TRACE_ACTOR: ci-bot

jobs:
  test-and-track:
    runs-on: ubuntu-latest
    
    steps:
      - uses: actions/checkout@v3
        with:
          fetch-depth: 0  # Need full history for git log
      
      - name: Install Rust
        uses: actions-rs/toolchain@v1
        with:
          toolchain: stable
      
      - name: Install Tracer
        run: cargo install tracer-cli
      
      - name: Initialize Tracer
        run: |
          tracer init
          git config --global user.email "[email protected]"
          git config --global user.name "CI Bot"
      
      - name: Run Tests
        id: tests
        run: |
          if cargo test 2>&1 | tee test_output.txt; then
            echo "status=success" >> $GITHUB_OUTPUT
          else
            echo "status=failure" >> $GITHUB_OUTPUT
          fi
      
      - name: Process Test Failures
        if: steps.tests.outputs.status == 'failure'
        run: |
          FAILED=$(grep "FAILED" test_output.txt | awk '{print $2}')
          
          for test in $FAILED; do
            TITLE="Test failure: $test"
            EXISTING=$(tracer list --json | jq -r ".[] | select(.title == \"$TITLE\" and .status != \"closed\") | .id" | head -n 1)
            
            if [ -z "$EXISTING" ]; then
              tracer create "$TITLE" -t bug -p 0
            else
              tracer comment $EXISTING "Still failing in build #${{ github.run_number }}"
            fi
          done
      
      - name: Update Issue Comments
        run: |
          ISSUES=$(git log -1 --pretty=%B | grep -oE 'bd-[0-9]+' | sort -u)
          BUILD_URL="${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
          
          for issue in $ISSUES; do
            if tracer show $issue --json > /dev/null 2>&1; then
              if [ "${{ steps.tests.outputs.status }}" = "success" ]; then
                tracer comment $issue "✓ Build #${{ github.run_number }} passed: $BUILD_URL"
              else
                tracer comment $issue "✗ Build #${{ github.run_number }} failed: $BUILD_URL"
              fi
            fi
          done
      
      - name: Export Tracer Database
        run: |
          tracer export
          git add .trace/
          git commit -m "Update Tracer state from build #${{ github.run_number }}" || true
      
      - name: Show Stats
        run: tracer stats

Environment Variables for CI

Common environment variable setup:

GitHub Actions
GitLab CI
Jenkins

env:
  TRACE_ACTOR: ci-bot
  TRACE_DB: .trace/issues.db
  BUILD_NUMBER: ${{ github.run_number }}
  BUILD_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}

variables:
  TRACE_ACTOR: ci-bot
  TRACE_DB: .trace/issues.db
  BUILD_NUMBER: $CI_PIPELINE_ID
  BUILD_URL: $CI_PIPELINE_URL

environment {
  TRACE_ACTOR = 'ci-bot'
  TRACE_DB = '.trace/issues.db'
  BUILD_NUMBER = "${env.BUILD_NUMBER}"
  BUILD_URL = "${env.BUILD_URL}"
}

Automation Best Practices

Use Dedicated CI Actor

# Always use a dedicated actor for CI
export TRACE_ACTOR="ci-bot"

# Not developer or agent names
# export TRACE_ACTOR="alice"  # Don't do this

Check Before Creating Issues

# Avoid duplicate issues
TITLE="Build failure: test_auth"
EXISTING=$(tracer list --json | jq -r ".[] | select(.title == \"$TITLE\" and .status != \"closed\") | .id" | head -n 1)

if [ -z "$EXISTING" ]; then
  tracer create "$TITLE" -t bug -p 0
else
  tracer comment $EXISTING "Still occurring in build #$BUILD_NUMBER"
fi

Include Build Context

# Always link back to CI build
BUILD_URL="https://ci.example.com/builds/$BUILD_NUMBER"
tracer comment $ISSUE "Build: $BUILD_URL"
tracer comment $ISSUE "Commit: $(git rev-parse --short HEAD)"
tracer comment $ISSUE "Branch: $(git branch --show-current)"

Auto-Close Resolved Issues

# Close issues when tests pass
if [ "$BUILD_STATUS" = "success" ]; then
  OPEN_FAILURES=$(tracer list --json | jq -r '.[] | select(.title | startswith("Test failure:")) | .id')
  
  for issue in $OPEN_FAILURES; do
    tracer close $issue --reason "Resolved in build #$BUILD_NUMBER"
  done
fi

AI Agents

Multi-Agent

Examples

Continuous Integration

CI/CD Benefits

Basic CI Integration

JSON Output Parsing

Test Failure Tracking

Build Status Comments

Agent Task Assignment from CI

Dependency Updates

Performance Regression Detection

Complete CI/CD Workflow

Environment Variables for CI

Automation Best Practices

Use Dedicated CI Actor

Check Before Creating Issues

Include Build Context

Auto-Close Resolved Issues

Next Steps

Example Scripts

Multi-Agent Coordination

Build docs developers (and LLMs) love

AI Agents

Multi-Agent

Examples

​CI/CD Benefits

​Basic CI Integration

​JSON Output Parsing

​Test Failure Tracking

​Build Status Comments

​Agent Task Assignment from CI

​Dependency Updates

​Performance Regression Detection

​Complete CI/CD Workflow

​Environment Variables for CI

​Automation Best Practices

​Use Dedicated CI Actor

​Check Before Creating Issues

​Include Build Context

​Auto-Close Resolved Issues

​Next Steps

Example Scripts

Multi-Agent Coordination

Build docs developers (and LLMs) love

CI/CD Benefits

Basic CI Integration

JSON Output Parsing

Test Failure Tracking

Build Status Comments

Agent Task Assignment from CI

Dependency Updates

Performance Regression Detection

Complete CI/CD Workflow

Environment Variables for CI

Automation Best Practices

Use Dedicated CI Actor

Check Before Creating Issues

Include Build Context

Auto-Close Resolved Issues

Next Steps