The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides specific commands for the agent to run, including the aqe CLI for quality assessments and a node -e one-liner to update a local history file. The Node.js command interacts directly with the file system using the fs module to read and write run-history.json.
[DYNAMIC_EXECUTION]: The instructions direct the agent to execute a JavaScript snippet at runtime using node -e. While this is intended for recording metrics, it represents a pattern of dynamic script execution provided within the skill's instructions.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing untrusted external data, such as source code in the src/ directory and test output. It does not employ boundary markers or sanitization to protect against malicious instructions embedded in that data. Furthermore, the agent's behavior is influenced by the contents of its own run-history.json file.
Ingestion points: Source code files in src/, test metrics, and run-history.json (SKILL.md).
Boundary markers: Absent; no specific delimiters are defined to separate untrusted data from instructions.
Capability inventory: Execution of shell commands (aqe) and Node.js code with file system read/write access (SKILL.md).
Sanitization: None; the skill interpolates and processes external code and metadata directly.
[TIME_DELAYED_CONDITIONAL]: The skill includes an instructional trigger that alters agent behavior (issuing an alert) based on historical state—specifically, if quality gates have failed in 3 of the last 5 recorded runs in run-history.json.

qe-quality-assessment