The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to utilize standard CLI tools like Bash and Grep for identifying security-sensitive patterns and performing dependency audits through commands like npm audit. these actions are legitimate and aligned with the skill's purpose.\n- [DATA_EXFILTRATION]: Instructions include searching for potentially hardcoded secrets (e.g., passwords or API keys) within the local source code using Grep. there is no evidence of network operations or intent to transmit this sensitive data to external entities.\n- [PROMPT_INJECTION]: As an auditing tool that processes untrusted project files, the skill has an inherent surface for indirect prompt injection. a maliciously crafted source file could attempt to influence the agent's behavior during the review process.\n
Ingestion points: The skill reads and greps various project files and configuration modules (SKILL.md, Steps 1-8).\n
Boundary markers: There are no explicit delimiters or instructions provided to separate the audited code content from the agent's instructions.\n
Capability inventory: Access to tools like Bash, Edit, Read, Grep, and Glob provides the agent with extensive interaction capabilities over the local environment.\n
Sanitization: The skill does not define specific sanitization or validation logic for the content it reviews.

typescript-security-review