ci-speedup

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill's phase‑4a gap‑fill explicitly requires the agent to read captured job logs and include verbatim log lines as "evidence" in the analysis JSON/report, which forces the LLM to handle and output raw log contents that may contain secrets or credentials.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). This skill’s required runtime workflow performs an LLM “gap-fill” (Phase 4a) by reading GitHub Actions job log text from outsider sources—collect_runs.py --with-logs fetches third-party job logs into data_bundle, and the agent running the skill then passes captured evidence/log lines into the LLM context.