The Agent Skills Directory

[DATA_EXFILTRATION]: The skill instructs the agent to access and verify sensitive environment configuration files which often contain secrets and credentials.
Evidence: In SKILL.md, the "Auto-debug steps" section explicitly directs the agent to "Verify .env.example vs actual config".
[COMMAND_EXECUTION]: The skill encourages the use of powerful shell commands to perform debugging tasks, which could be exploited if the agent is misled.
Evidence: In SKILL.md, the agent is instructed to use bash to "run dev server, test suite, or reproduce" and to use git log and grep across the codebase.
[PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection by instructing the agent to process data from potentially untrusted sources.
Ingestion points: The agent is directed to read codebase files, search results, git history, and server/error logs in SKILL.md. Additionally, DEBUG-PROMPTS.md contains templates for users to provide console and network log data.
Boundary markers: There are no instructions or delimiters provided to ensure the agent ignores or sanitizes instructions that might be embedded within the diagnostic data or log files.
Capability inventory: The skill utilizes file system access, shell command execution (bash), and git operations.
Sanitization: No sanitization or validation mechanisms are described for the external content before it is interpolated into the agent's context.

debug