forecast-ledger
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a Python script (
ledger.py) to manage forecast data. It performs standard file I/O on a JSONL file to persist forecast entries. - [DATA_EXPOSURE]: Data is stored locally in
.cache/forecast-ledger/ledger.jsonlby default, or at a path specified by theFORECAST_LEDGERenvironment variable. The skill only processes forecast-related data (probabilities, dates, questions) and does not access sensitive system files or user credentials. - [REMOTE_CODE_EXECUTION]: The script uses standard libraries (
argparse,json,os,sys,datetime) and does not perform any network operations, dynamic code execution (eval/exec), or external script downloads. - [PROMPT_INJECTION]: The instructions in
SKILL.mdare descriptive and provide clear guidance on how to use the CLI tool. No attempts to override agent behavior or bypass safety filters were found.
Audit Metadata