The Agent Skills Directory

[COMMAND_EXECUTION]: The skill automatically detects and executes arbitrary commands derived from project files such as Makefile, justfile, Taskfile.yml, and package.json. It also searches for and executes scripts within bin/ and .my/bin/ directories, allowing for the execution of code found within the local environment.
[COMMAND_EXECUTION]: Executes the jj diff command to interact with the local Jujutsu version control system.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes instructions from multiple untrusted sources to determine its actions.
Ingestion points: The skill reads project documentation including README.md, CONTRIBUTING.md, CLAUDE.md, AGENTS.md, GEMINI.md, and CODEX.md (SKILL.md).
Boundary markers: Absent. The skill does not use delimiters or instructions to ignore embedded prompts when reading these files.
Capability inventory: The skill has the capability to execute shell commands and scripts discovered during the detection phase (SKILL.md).
Sanitization: Absent. There is no validation or filtering of the commands or instructions retrieved from the processed files.

code-test