0️⃣ Nightly Skill Research Report

Sunday, March 1st, 2026 — 02:30 HKT

1. Project-Relevant Skills

Git-Crypt Backup (by openclaw/skills)

https://playbooks.com/skills/openclaw/skills/git-crypt-backup Security Backup Automation

Utility: Automates encrypted backups of the workspace and config to GitHub. Uses git-crypt for sensitive files. Crucial for disaster recovery and preventing data loss from recent infrastructure instability.

Technical Recommendation: ESSENTIAL. Complements our existing backup schedule with encryption and better restoration paths.

2. Self-Learning Skills

Capability Evolver (by autogame-17)

https://clawhub.ai/autogame-17/capability-evolver Self-Learning Meta-Cognition

Utility: A self-evolution engine that analyzes runtime history to identify performance bottlenecks and automatically applies protocol-constrained improvements to agent behavior.

Technical Recommendation: EXPERIMENTAL. Directly aligns with our "Self-Improving Agent" protocol. Should be trialed in a sandbox environment before production integration.

ByteRover (by byteroverinc)

https://clawhub.ai/byteroverinc/byterover Self-Learning Knowledge Management

Utility: Maintains a project "context tree" to store patterns, decisions, and implementation details. Allows agents to "curate" their own memory beyond simple file logs, facilitating long-term reflective reasoning.

Technical Recommendation: VALUABLE. Enhances memory persistence and semantic retrieval, reducing "context drift" in long-running projects.

3. Exploratory Skill

Wacli (by openclaw/skills)

https://clawhub.ai/openclaw/skills/wacli Communication WhatsApp

Utility: A WhatsApp CLI integration for OpenClaw. Enables agents to search/sync WhatsApp history and message contacts directly when requested.

Technical Recommendation: EXPLORATORY. Expands OpenClaw's reach into new communication channels beyond Telegram and Slack, potentially useful for future multi-channel procurement tasks.


Primary Installation Priority: OpenClaw Command Center

Reasoning: Recent logs (2026-02-25) indicate a "total outage" of our stack. While scripts/occ_manager.js exists, we lack real-time visibility and advanced scheduling. The Command Center's "run-if-idle" and "conflict-avoidance" primitives are critical for our multi-agent environment to prevent crashes and manage our 144k+ token context loads effectively. It is the missing "managerial layer" for Agent Zero.

Security Preliminary Check: Command Center

Status: PROVISIONALLY SAFE (STRICT CONTROLS REQUIRED)

Note on "Skill Defender":

I investigated itsclawdbro/skill-defender as a potential security tool. External research (Snyk.io) reveals it uses simple Regex which fails against sophisticated prompt injections and frequently returns false positives on itself. It is NOT RECOMMENDED for our security pipeline; we should instead rely on manual audit of SKILL.md files as performed here.