The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill implements a dynamic code execution pattern in Step 5. It uses a shell command (python3 -c) to run a script that leverages importlib.util.spec_from_file_location and spec.loader.exec_module to load and execute classes from a Python file at a path determined at runtime. This is used to smoke-test the solution code against example inputs.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it ingests and processes untrusted data from external sources to generate executable code.
Ingestion points: Data is read from contest scratch files (leetcode_python/lc_weekly/*/ws.txt) and user-pasted solution drafts.
Boundary markers: The instructions do not define specific delimiters or guardrails to prevent the agent from following instructions potentially embedded within these external inputs.
Capability inventory: The skill has the ability to write files to the local filesystem and execute code via the shell, which could be exploited if malicious content is present in the ingested data.
Sanitization: There is no evidence of sanitization or validation of the input data before it is written to the solution files and subsequently executed during the test phase.
[COMMAND_EXECUTION]: The skill utilizes several standard shell utilities to automate repository tasks. This includes grep and sed for searching contest notes, ls and git log for directory inspection, and awk for verifying the structure of the README.md file after updates.

lc-python