coding-prompt
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill is a coherent coding-prompt coach, with minor notes because it can stay active during a session and optionally append learned preferences to a local reference file.
This skill appears safe for its stated purpose. Before using the update-skill feature, review what it wants to append to learnings.md and avoid saving confidential project details or personal information.
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.
After you activate it, the agent may keep adding prompt-quality warnings during the same session.
This shows session-level persistence and proactive behavior after activation, but it is limited to two coding-prompt signals and is not hidden.
Once activated (Mode 1 triggered), the skill remains active for the rest of the session... proactively alert
Use activation intentionally and ask the agent to stop monitoring if the ongoing alerts are not wanted.
Session preferences or lessons could be reused in future prompt optimizations, including details the user may not want persisted.
The skill can persist reusable learnings and preferences from coding sessions into a local reference file when the update protocol is used.
`references/learnings.md` | ✅ **APPEND-ONLY** | Personal experience layer — the sole evolution target
Review proposed learnings before appending them, keep entries generic, and avoid storing secrets, proprietary project details, or one-off context.