The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection (Category 8) as it automatically ingests data from external project files to determine its naming strategies and context.
Ingestion points: Processes content from README, package.json, documentation files, and various configuration files during the 'Context Discovery' phase (SKILL.md).
Boundary markers: Absent. The instructions do not define delimiters or provide warnings to the agent to ignore potential instructions embedded within the files it scans.
Capability inventory: The skill allows the use of web searches, registry checks (npm, GitHub), and any other available tools to perform research and validation.
Sanitization: Absent. No filtering or validation is performed on the data read from local files before it is used to influence the agent's output.
[DATA_EXFILTRATION]: The skill combines local file access with network operations, creating a potential path for data exposure.
Sensitive Access: The agent is instructed to scan config files for descriptive metadata. Depending on the environment, this could lead to the inspection of sensitive configuration or environment data.
Network Operations: The skill performs a "silent research pass" and "Collision Clearance" checks against external domains (npm, GitHub, and general web searches) using information derived from the project context.

name-project