The Agent Skills Directory

[COMMAND_EXECUTION]: The skill generates and encourages the execution of shell scripts for application booting, health checks, and pre-push enforcement (e.g., scripts/init.sh, worktree-start.sh, and .git-hooks/pre-push). These scripts use standard tools like curl, jq, and package managers to verify system state.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it processes and analyzes content from the repository it is auditing.
Ingestion points: The skill reads files from the repository filesystem (source code, READMEs, configuration files) to perform the readiness audit described in SKILL.md.
Boundary markers: There are no specified boundary markers or instructions to disregard potential injection content within the files being audited.
Capability inventory: The agent has the ability to create executable scripts, write to the filesystem, and run shell commands, which could be leveraged if it follows malicious instructions in the target repository.
Sanitization: The skill does not implement sanitization or validation of the text ingested from repository files before it enters the agent's context.
[SAFE]: All external URL references point to official domains and repositories of well-known technology organizations such as OpenAI, Anthropic, Stripe, Datadog, and Meta.
[SAFE]: No obfuscation, multi-layer encoding, or unauthorized access to sensitive credentials or configuration files (like .env or .ssh) was detected.

agent-readiness