The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted git diff data which could contain malicious instructions intended to influence the agent's risk assessment, classification, or reviewer recommendations.
Ingestion points: Git diffs and pull request data are ingested at runtime via tools like mcp__plugin_ruflo-core_ruflo__analyze_diff and Bash commands (e.g., gh pr view) as described in Steps 0 and 1 of SKILL.md.
Boundary markers: The instructions do not specify any boundary markers or delimiters to separate the diff content from the agent's instructions, nor do they include warnings to ignore instructions embedded within the processed data.
Capability inventory: The skill utilizes a suite of analysis tools (mcp__plugin_ruflo-core_ruflo__analyze_diff, analyze_diff-risk, etc.) and the Bash tool to interact with the file system and external Git repositories.
Sanitization: No evidence of sanitization or filtering of the diff content is present in the skill instructions to mitigate the impact of adversarial content within the code changes.

diff-analyze