catchup
Warn
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the 'Bash' tool to execute shell commands where user-provided input from the '$ARGUMENTS' variable is interpolated without quoting or sanitization. In 'SKILL.md', the variable '$BASE_BRANCH' (derived from '$ARGUMENTS') is used in several commands such as 'git show-ref --verify --quiet refs/heads/$BASE_BRANCH', 'git rev-list --count $BASE_BRANCH..HEAD', and 'git diff --stat $BASE_BRANCH...HEAD'. This allows arbitrary command execution if a user provides input containing shell metacharacters like semicolons, backticks, or dollar-parenthesis.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from the user via the '$ARGUMENTS' variable and uses it to influence powerful tool capabilities (Bash) without using boundary markers or sanitization. Evidence: 1. Ingestion points: '$ARGUMENTS' in 'SKILL.md'. 2. Boundary markers: Absent. 3. Capability inventory: 'Bash' tool executing git commands in 'SKILL.md'. 4. Sanitization: Absent.
Audit Metadata