OpenClaw Growth Pack
ReviewAudited by ClawScan on May 10, 2026.
Overview
This instruction-only skill is mostly transparent, but it asks to change core OpenClaw configuration, handle gateway/API tokens, and add persistent autonomous self-check behavior that should be reviewed before use.
Install only if you intentionally want this OpenClaw instance to gain persistent anti-stall and autonomy behavior. Before applying it, confirm which files will be edited, redact tokens during audits, avoid storing secrets in memory files, and require an explicit rollback plan for AGENTS.md, HEARTBEAT.md, memory rules, and any cron/system jobs.
Findings (4)
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.
The agent could keep checking or acting on tasks on a schedule after the initial setup, which may surprise users if not explicitly approved and documented.
The skill instructs creation of recurring background behavior that can continue after the setup task, but it does not specify exact job scope, approval requirements, disable steps, or cleanup.
If cron/system events are available, create conservative jobs: - Daily: unfinished-task check. - Weekly: memory review and friction pattern extraction.
Only enable scheduled jobs with explicit user opt-in; document the exact command, schedule, working directory, logs, and a complete disable/removal procedure.
Future agent sessions may follow these anti-stall rules even when a different project or user preference would call for more cautious stopping or confirmation.
The skill persists broad mandatory agent behavior and stopping-condition instructions, which may affect future tasks beyond the immediate setup.
Write or update `AGENTS.md` with these mandatory constraints: - Output state on each substantial task: `Goal`, `Progress`, `Next`. - Do not stop before completion except for explicit blocker or user stop.
Scope AGENTS.md changes to the intended workspace, make them opt-in, preserve user stop/approval requirements for risky actions, and provide a rollback for all inserted instructions.
Gateway tokens or provider keys could be accidentally revealed to anyone with access to logs or the agent conversation.
The PowerShell audit reads gateway tokens from local configuration and prints their full values, which can expose credentials in terminal logs, transcripts, or chat history.
"auth.token = $auth" "remote.token = $remote"
Do not print full tokens. Compare token presence or hashes, show only a short redacted prefix/suffix, and clearly warn users before handling credential files.
Incorrect or sensitive information written to memory files may be reused in later tasks or influence future agent behavior.
The skill uses persistent memory files and later review of those files to influence future behavior; this is purpose-aligned but should be treated as persistent agent context.
If queue item exists, execute one concrete step, then log evidence to `memory/YYYY-MM-DD.md`.
Review memory files regularly, avoid storing secrets, and provide a way to prune or reset memory and derived local rules.