The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): The skill identifies and executes scripts directly from the package.json file (e.g., build, type-check, lint). Because it blindly trusts the contents of this configuration file, a malicious repository could execute arbitrary shell commands on the host system by defining them within these script tags.
[PROMPT_INJECTION] (HIGH): This skill is highly susceptible to Indirect Prompt Injection (Category 8).
Ingestion points: Reads package.json, project source files, and ESLint diagnostic output.
Boundary markers: Absent. The workflow does not include instructions to ignore or delimit embedded natural language instructions within the code.
Capability inventory: Executes shell commands (npm, python3), performs file modifications, and executes git commit operations.
Sanitization: Absent. The skill does not filter or sanitize the content of the files it processes before making decisions or executing fixes.
Risk: An attacker could place malicious instructions in code comments or strings that trick the agent into exfiltrating sensitive environment variables or introducing backdoors during the 'fixing' process.
[REMOTE_CODE_EXECUTION] (MEDIUM): The skill references a local Python script scripts/sort_eslint_by_rule.py which is not provided in the skill package. If an attacker can place a file at this path, the agent will execute it using python3.

lint-build-fixer