JUnit Failure Fingerprint

Cluster JUnit failures into stable fingerprints so CI triage focuses on root causes, not noisy one-off logs.

Security Scan

The skill is internally consistent: it only parses local JUnit XML files, normalizes noisy tokens, and emits clustered fingerprints; it does not request credentials, external network access, or install software.

Purpose & Capability

Name and description (cluster JUnit failures) match the included script and SKILL.md. Required binaries (bash, python3) are appropriate and sufficient for the stated task; there are no unrelated credentials or config paths.

Instruction Scope

Runtime instructions and the script limit actions to scanning JUnit XML files (via a glob), normalizing text, hashing seeds, and printing/grouping results. Caution: JUNIT_GLOB is user-controllable — if set to a broad or sensitive filesystem path, the script will read whatever files match the glob. By default the glob targets test-results/**/*.xml and the bundled fixtures example is safe.

Install Mechanism

No install spec; this is instruction-only plus a bundled script. Nothing is downloaded or written to disk by an installer during skill installation.

Credentials

The skill requests no environment variables or credentials by default. It does accept optional environment inputs (JUNIT_GLOB, TOP_N, OUTPUT_FORMAT, STACK_LINES, FAIL_ON_FAILURES) that are reasonable for its purpose.

Persistence & Privilege

Skill is not always-enabled and does not request elevated persistence or modify other skills or system settings. It only runs when invoked and doesn't persist state or credentials.

Assessment

This skill appears safe and does exactly what it claims: parse JUnit XMLs and produce fingerprints. Before using in CI, verify your JUNIT_GLOB value so it only matches test result files (not / or other sensitive locations). You can test locally with the provided fixtures. If you enable FAIL_ON_FAILURES, be aware it will return a non-zero exit code when failures are found, which can fail CI jobs as intended.

Use this skill to compress noisy JUnit failures/errors into repeatable fingerprints.

What this skill does

• Scans one or more JUnit XML files
• Extracts only failing/error test cases
• Normalizes volatile tokens (IDs, numbers, line numbers, addresses, UUIDs)
• Generates stable fingerprint hashes for similar root-cause failures
• Emits grouped triage output (text or json)

Inputs

Optional:

• JUNIT_GLOB (default: test-results/**/*.xml)
• TOP_N (default: 20)
• OUTPUT_FORMAT (text or json, default: text)
• STACK_LINES (default: 3) — number of normalized stack lines to include in fingerprint seed
• FAIL_ON_FAILURES (0 or 1, default: 0) — exit non-zero when any failures/errors are found

Run

Text report:

JUNIT_GLOB='artifacts/junit/**/*.xml' \
TOP_N=15 \
bash skills/junit-failure-fingerprint/scripts/fingerprint-junit.sh

JSON output for CI annotation/upload:

JUNIT_GLOB='artifacts/junit/**/*.xml' \
OUTPUT_FORMAT=json \
FAIL_ON_FAILURES=1 \
bash skills/junit-failure-fingerprint/scripts/fingerprint-junit.sh

Run with bundled fixtures:

JUNIT_GLOB='skills/junit-failure-fingerprint/fixtures/*.xml' \
bash skills/junit-failure-fingerprint/scripts/fingerprint-junit.sh

Output contract

• Exit 0 by default (reporting mode)
• Exit 1 if FAIL_ON_FAILURES=1 and at least one failure/error is found
• In text mode, prints summary + top fingerprints
• In json mode, prints machine-readable groups and per-case detail