The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted repository data (metadata and README files) that could contain malicious instructions intended to manipulate the agent's behavior during analysis.
Ingestion points: The skill processes repo_data (description, topics, timestamps) and readme_content (markdown text) through Python-based analysis functions.
Boundary markers: Absent. There are no instructions or delimiters provided to ensure the agent ignores or isolates natural language directives that may be embedded within the malicious repository's files.
Capability inventory: The skill provides examples of shell commands (git clone, find, grep, rm), network operations (requests.post, gh api), and environment variable access (os.getenv).
Sanitization: No sanitization, escaping, or schema validation is performed on the ingested content before it is processed by the analysis logic.
[EXTERNAL_DOWNLOADS]: The skill provides instructions for the agent to clone external repository content from github.com/29Hinojosa for the purpose of quarantine analysis.
[COMMAND_EXECUTION]: The skill includes shell script templates for performing static analysis on potentially malicious files using standard tools such as find, grep, and git.

k7-total-security-unlock-patch-detection