Inner Life Evolve
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill is coherent with its stated purpose: it reads the agent’s local memory/state and writes proposed improvements to a task queue for user approval, without evidence of hidden execution or data exfiltration.
Before installing, understand that this skill reads your agent’s local memory/state and writes self-improvement suggestions into a task queue. It appears designed for human approval rather than automatic changes, so review any `[EVOLVER]` proposals carefully before accepting them.
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.
If the stored memory or state is inaccurate or manipulated, the skill may produce poor or misleading improvement proposals.
The skill intentionally reads persistent agent memory and state to generate improvement proposals, which is aligned with its purpose but means future suggestions depend on stored context quality and trustworthiness.
Read everything: - AGENTS.md, TOOLS.md, BRAIN.md, SELF.md - `memory/week-digest.md` - `memory/habits.json` - `memory/drive.json` - `memory/relationship.json` - `memory/inner-state.json`
Review the generated proposals before approving them, and keep the underlying memory files accurate and free of untrusted instructions.
Unreviewed or low-quality proposals could keep reappearing and influence future decisions if the user accepts them without scrutiny.
The skill writes proposed behavior changes into a persistent queue that another workflow may later surface, so a bad proposal could persist across sessions even though execution is intended to require user approval.
Write proposals to `tasks/QUEUE.md` under the Ready section ... Brain Loop reads QUEUE and shows `[EVOLVER]` tasks at lower priority
Treat `[EVOLVER]` tasks as suggestions only, and approve them only after checking that they are specific, safe, and aligned with your goals.