fix-bug

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill processes external log data from the 'loki-runner' agent to automatically drive database queries, which introduces a potential surface for indirect prompt injection. * Ingestion points: Log data retrieved from the 'loki-runner' sub-agent (SKILL.md). * Boundary markers: No explicit sanitization or delimiters are defined for parsing table names and IDs from logs in the Stage 2 logic. * Capability inventory: Database access via 'mysql-runner', code analysis via 'bug-analyzer', and file modification/git commits via 'git-workflow'. * Sanitization: The skill mandates a 'Bug 排查报告' (Bug Investigation Report) that must be manually confirmed by the user before any destructive actions or commits are executed.
  • [COMMAND_EXECUTION]: The skill orchestrates the execution of SQL commands and repository operations via sub-agents. It dynamically generates database queries based on information extracted from logs and bug descriptions, and uses the 'git-workflow' skill to perform final code submissions.
  • [CREDENTIALS_UNSAFE]: The skill interacts with configuration files that store credentials for database and log systems. * Evidence: References to '.claude/mysql-config.json' and '.claude/skills/loki-log-query/environments.json'. * Mitigation: The skill includes explicit instructions forbidding the AI from writing passwords or tokens from the conversation into these files, requiring manual user configuration.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 10:11 AM