Crypto Self-Learning
ReviewAudited by ClawScan on May 10, 2026.
Overview
This looks like a local trade journaling and analysis skill, but it stores trading history and can persist generated rules into agent memory, so review changes before relying on them.
This skill appears safe to use as a local trade log and analysis tool. Before installing, be aware that it keeps financial trading records in local JSON files and can write learned rules into agent memory; preview memory updates with --dry-run and do not rely on its rules as financial advice.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Incorrect, stale, or overfit trade history could become persistent guidance and bias future crypto-trading recommendations.
The script can write generated trading rules into an existing agent memory file chosen by path, creating persistent context that future agent sessions may use.
parser.add_argument("--memory-path", required=True, help="Path to MEMORY.md") ... with open(memory_path, "w") as f: f.write(new_content)Use --dry-run first, review the generated rules and trade history, back up MEMORY.md, and only write to the intended memory file.
The advertised weekly review command may fail, or a user might be tempted to obtain and run an unreviewed helper from elsewhere.
The documentation instructs users to run a weekly_review.py helper, but that file is not present in the supplied manifest or file contents.
python3 {baseDir}/scripts/weekly_review.pyDo not run missing helper scripts from outside the package unless they are reviewed; the maintainer should include the file or remove the command.
A user or agent may apply learned rules too rigidly and make poor trading decisions based on limited or non-representative data.
The wording encourages strong reliance on generated rules in a financially sensitive trading context.
Trust the data - If data says avoid something, AVOID IT
Treat the output as journaling and analysis, not guaranteed financial advice; require human review before making trades.