The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The SKILL.md file uses the !command`` syntax to execute shell commands automatically at load time. These commands are intended to validate the local dev toolchain by checking for the presence of mise, fnox, and infisical.
[DYNAMIC_CONTEXT_INJECTION]: The load-time commands retrieve identity information by executing infisical user get, which exposes account details (such as email or user ID) to the agent upon loading.
[DYNAMIC_CONTEXT_INJECTION]: The skill automatically reads system configuration by executing head on files in the user's home directory, specifically ~/.config/mise/config.toml and ~/.config/fnox/config.toml, bringing local environment settings into the agent's context.
[EXTERNAL_DOWNLOADS]: The skill documentation suggests installing developer tools via remote scripts (e.g., curl | sh). The download sources (mise.run and dl.infisical.com) are the official distribution points for these well-known services.
[COMMAND_EXECUTION]: The reference files provide documentation and examples for executing shell commands with injected secrets using the fnox exec and infisical run tools.
[INDIRECT_PROMPT_INJECTION]: The skill ingests local configuration data from the current directory at load time.
Ingestion points: Project files fnox.toml, .infisical.json, and mise.toml are read via !command`` in SKILL.md.
Boundary markers: There are no explicit boundary markers or instructions to ignore embedded commands when reading these files.
Capability inventory: The skill documents the ability to execute shell commands using the integrated secret management toolchain.
Sanitization: No sanitization is performed on the content read from project files before it is processed by the agent.

devtools-secrets