The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The setup.sh script and references/lean.md documentation execute the elan-init.sh installation script from elan-lang.org by piping the download directly to the shell (sh). This follows the standard, official installation procedure for the Lean toolchain manager.
[EXTERNAL_DOWNLOADS]: The skill downloads required binaries and scripts from official sources, including tla2tools.jar from the TLA+ GitHub releases and the elan installer from the Lean project's official domain.
[COMMAND_EXECUTION]: The skill utilizes local shell commands to build models and run model checkers, specifically using lake build for Lean 4 projects and java for the TLA+ TLC model checker.
[INDIRECT_PROMPT_INJECTION]: The skill performs analysis on repository code, which serves as a potential vector for indirect prompt injection.
Ingestion points: The agent is instructed to read source code, tests, and comments from the target repository in the SKILL.md file (Steps 1 and 2).
Boundary markers: There are no specific instructions to use delimiters or ignore embedded instructions when reading repository content.
Capability inventory: The agent writes formal model files to the filesystem and executes them using the Lean and TLA+ toolchains.
Sanitization: The skill lacks explicit sanitization or validation of the content extracted from the repository before it is used to generate executable models.

lean