The Agent Skills Directory

[SAFE]: No malicious code, obfuscation, or data exfiltration patterns were detected. The skill utilizes standard local tools (ls, cp, git, Read, Write) to assist in understanding project structures and data pipelines.
[SAFE]: The skill mandates strict data preservation guidelines, including a "CRITICAL RULE" to never delete files and instructions to use a legacy directory for safekeeping original code and data.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as its primary function is to ingest and analyze untrusted files from a user-specified path.
Ingestion points: The skill reads arbitrary content from files in the provided project-path using tools like Read, Grep, and Glob.
Boundary markers: No explicit delimiters are specified in the prompt templates to isolate the ingested file content from the agent's operational instructions.
Capability inventory: The skill is granted file system write and edit permissions (Write, Edit, Bash(cp*)), which could be targeted if a file contains malicious instructions.
Sanitization: Content from analyzed files is processed directly without filtering or sanitization.
Note: This risk is inherent to the auditing use case and is mitigated by the skill's explicit safety instructions and the agent's core safety guardrails.

code-archaeology