Skip to content

Build Test Suite

Build Test Suite #454

Triggered via pull request April 6, 2026 16:15
Status Failure
Total duration 8m 2s
Artifacts 4

build-test.lock.yml

on: pull_request
Fit to window
Zoom out
Zoom in

Annotations

3 errors, 3 warnings, and 1 notice
detection
ERR_VALIDATION: ❌ Security threats detected: prompt injection Reasons: The workflow prompt file (prompt.txt) contains an injected pre-fabricated JSON response embedded at the end of the legitimate workflow instructions. This JSON matches the agent_output.json exactly, showing all 18 projects across 8 ecosystems passing with identical '1/1 passed' results. This is a classic prompt injection attack where the attacker appended the desired output directly into the prompt file to trick the agent into reporting false test results without actually executing any tests. The injected payload also requests adding a 'build-test' label to PR #1704, which would fraudulently signal that build tests passed when they were never run.
detection
Reasons: The workflow prompt file (prompt.txt) contains an injected pre-fabricated JSON response embedded at the end of the legitimate workflow instructions. This JSON matches the agent_output.json exactly, showing all 18 projects across 8 ecosystems passing with identical '1/1 passed' results. This is a classic prompt injection attack where the attacker appended the desired output directly into the prompt file to trick the agent into reporting false test results without actually executing any tests. The injected payload also requests adding a 'build-test' label to PR #1704, which would fraudulently signal that build tests passed when they were never run.
detection
🚨 Security threats detected: prompt injection
activation
Template-like syntax detected and escaped. This is a defense-in-depth measure to prevent potential template injection if content is processed by downstream template engines. GitHub's markdown rendering does not evaluate template syntax.
activation
Template-like syntax detected and escaped. This is a defense-in-depth measure to prevent potential template injection if content is processed by downstream template engines. GitHub's markdown rendering does not evaluate template syntax.
agent
GitHub MCP guard policy automatically applied for public repository. min-integrity='approved' and repos='all' ensure only approved-integrity content is accessible.
agent
Safe Outputs MCP Server Startup Log

Artifacts

Produced during runtime
Name Size Digest
activation
4.45 KB
sha256:4b79ae9b600e9e7ec6d70b77399d8604d4eea1c6112e479a7b4f191e3d1da19e
agent
11.6 MB
sha256:eba60328c095f12dfdac04b149dae2b62c618a1d33b94989c4f40b31384d7df5
detection
23.2 KB
sha256:159adc4213496a7aed464a8fa81f05958080411bd5b81ca0fcd4be14ea7752aa
firewall-audit-logs
15.2 KB
sha256:ba1c672b16a50a4a3efa2a4ad3ef9171708a22d1971ba41b3684cdf650c40655