The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from multiple external sources that could contain malicious instructions designed to influence the agent's behavior or reporting.
Ingestion points: The skill reads aeon.yml, skills/*/SKILL.md, and memory logs from memory/logs/ (SKILL.md).
Capability inventory: The skill uses grep, sort, head, tail, wc, cat, jq, node for analysis, and ./notify for reporting (SKILL.md).
Boundary markers: The skill does not use explicit boundary markers to delimit untrusted content when processing file data.
Sanitization: The skill uses regular expressions to extract specific fields and includes some basic string escaping in its Node.js scripts, but it lacks comprehensive sanitization for the data passed to the notification tool.
[DYNAMIC_EXECUTION]: The skill uses inline Node.js scripts to perform complex validation logic at runtime.
Evidence: Multiple node -e commands are defined in the instruction steps to parse YAML frontmatter and file system structures (Steps 2, 3, 5, 11, and 12 in SKILL.md).
[COMMAND_EXECUTION]: The skill executes shell commands and a vendor-specific notification utility.
Evidence: The skill uses grep for pattern matching and calls ./notify to send consolidated findings to the operator (Steps 1, 6, 7, 8, 9, 10, and the Report section in SKILL.md).

aeon-doctor