The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute a local utility script (estimate-tokens.sh) for token estimation and recommends running standard security scanners such as npm audit, pip-audit, and cargo audit to identify vulnerabilities in dependencies.
[PROMPT_INJECTION]: As the skill is designed to load an entire codebase (up to 125K lines of code) into the agent's context for analysis, it is inherently vulnerable to indirect prompt injection. If the audited codebase contains malicious instructions or adversarial patterns embedded in comments or strings, it could manipulate the agent's reporting or behavior.
Ingestion points: Files are loaded using Read, Glob, and Grep across the entire project directory.
Boundary markers: The skill does not explicitly define boundary markers or 'ignore' instructions for the ingested code content.
Capability inventory: The agent has access to Bash, Read, TaskCreate, and PushNotification tools.
Sanitization: No evidence of sanitization or escaping of the ingested source code before it enters the prompt context.
[DATA_EXPOSURE]: A hardcoded absolute file path is present in SKILL.md (/Users/yonatangross/coding/yonatangross/orchestkit/plugins/ork/skills/chain-patterns/rules/push-notification-on-completion.md). This discloses information about the author's local file system structure and username.

audit-full