The Agent Skills Directory

[PROMPT_INJECTION]: The instructions in instructions.md are focused on construction project management tasks and do not contain any instructions to override safety filters or ignore previous rules.
[DATA_EXFILTRATION]: There are no hardcoded credentials, sensitive file paths (like .ssh or .aws), or network operations (curl, wget, fetch) found in the provided code. The skill operates locally on the project data provided by the user.
[REMOTE_CODE_EXECUTION]: The skill uses standard Python libraries (pandas, numpy) and does not perform any remote script execution or download code from external URLs.
[OBFUSCATION]: The code and documentation are written in clear, plain text with no use of Base64 encoding, zero-width characters, or hidden text patterns.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to ingest external data (CSV, Excel, JSON), it processes this data through structured Python classes for mathematical optimization. There is no evidence that it interprets this data as instructions, though as with any data-processing skill, standard caution with untrusted input files is recommended.
[COMMAND_EXECUTION]: The skill does not spawn subprocesses or execute arbitrary shell commands. It performs all calculations within the Python runtime using established data science libraries.

resource-leveler