Firstprinciples thinking
PassAudited by ClawScan on May 1, 2026.
Overview
This is a coherent local reasoning skill, with the main thing to notice being that its optional helper scripts can save problem analyses locally for later reuse.
This skill looks safe for normal reasoning use. If you run its helper scripts, remember that your problem statements and derived analyses can be saved locally under the OpenClaw workspace, so avoid secrets or delete the saved memory if you do not want long-term retention.
Findings (2)
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.
A user might expect only prompt instructions, but the package also includes optional local scripts.
The metadata frames the skill as having no install mechanism while the package still contains helper code. The code is disclosed and appears benign, but the mismatch is worth noticing before invoking any scripts.
No install spec — this is an instruction-only skill. ... 13 code file(s): scripts/analyze_problem.py ... scripts/lib/storage.py
Treat the scripts as optional helper utilities; review them before running and prefer using the SKILL.md-only behavior if you do not need saved local cases.
If you run the helper scripts with confidential strategy, business, product, or personal problems, those details may remain in local saved-case files.
The analysis helper stores the user-provided problem text as part of a saved case, creating persistent local memory that can later be listed, shown, scored, exported, or used for promoted patterns.
"problem": args.text.strip(), ... data["cases"][case_id] = case
save_cases(data)Avoid entering secrets or highly sensitive details into saved cases unless local persistence is acceptable; periodically review or delete the local firstprinciples memory folder if needed.