Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Threat detection fails the run when it was skipped and `continue-on-error: false`

Open Beginner friendly
  • Is something wrong?
#66,872 0 comments 1 reaction 2 assignees View on GitHub

Maintainers usually reply within 1 day

@pelikhan is already working on this.

Since Oct 11, 2026.

  • #67810 by @copilot-swe-agent — open

Assessment

Difficulty
2/5
Estimated time
1-3 hours
Newbie friendliness
76/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
github-actions, go, yaml
Domain
ci-cd, devtools

Research direction

Start by reading conclude_threat_detection.sh and the detection job's guard and install steps; reproduce with continue-on-error: false and no safe outputs. Check how the conclusion handles RUN_DETECTION: false when installation was skipped. Done when detection reports skipped without failing in that case, while a missing detector still errors when detection was expected to run.

Written by the indexing model from the issue text.

Description

ai-inspected community

Problem

With safe-outputs.threat-detection.continue-on-error: false, a run in which the agent emits no safe output at all ends with a failed detection job:

ERR_SYSTEM: threat-detect binary not found on PATH
##[error]Process completed with exit code 1.

Nothing was there to analyze. The guard step says so ("Detection skipped: no agent outputs or patches to analyze"), and the run fails anyway, with a message that points at the detector install rather than at the agent.

This happens every time the agent stops before writing to outputs.jsonl. In our case the agent hit max-ai-credits (403 Maximum AI credits exceeded), so outputs.jsonl was never created, and the only failure shown on the run was the threat-detect one.

Seen on v0.89.21 and v0.91.5.

Repro

  1. A workflow with any safe output and: 
    safe-outputs:
  threat-detection:
    continue-on-error: false
  2. Make the agent end without calling any safe-output tool. For instance, set a max-ai-credits low enough to stop it early, or use a prompt that ends without calling any tool, noop included.
  3. Look at the detection job:
    • Check if detection needed: OUTPUT_TYPES: empty, HAS_PATCH: false, "Detection skipped: no agent outputs or patches to analyze".
    • Install threat-detect binary: skipped.
    • Conclude threat detection: env has RUN_DETECTION: false, THREAT_DETECT_INSTALL_OUTCOME: skipped, DETECTION_AGENTIC_EXECUTION_OUTCOME: skipped, then ERR_SYSTEM: threat-detect binary not found on PATH and exit code 1.

Expected

When the guard skipped detection because there was nothing to analyze, the conclusion should report it as skipped, as the comment in conclude_threat_detection.sh describes, whatever continue-on-error is set to. A missing detector should only be reported as an error when detection was actually expected to run.

Dominant language
Go
Stars
5.4k
Forks
576
Avg merge
8h 22m
Merged PRs (30d)
798

Getting set up

Open in Codespaces

Starts the project's dev container in your browser, under your own GitHub account.

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from github/gh-aw

All issues in github/gh-aw

Similar issues

More Go issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.