Workspace Review
v1.0.2Audit workspace structure and memory files against OpenClaw conventions. Use when asked to "review workspace", "audit files", "check structure", or during periodic self-maintenance. Helps catch drift from standard patterns.
⭐ 4· 2.7k·10 current·10 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description ('Workspace Review') match the provided SKILL.md, references, and the included shell script. All requested artifacts (checks, file locations, git status, memory hygiene) are directly relevant to auditing an OpenClaw workspace. No unrelated credentials, binaries, or config paths are required.
Instruction Scope
SKILL.md instructs inspection actions (ls, wc, git status, skimming files) appropriate for an audit. It documents platform behaviors (automatic memory flush / system prompt used for compaction) — informational for operators but flagged by the pre-scan as a possible 'system-prompt-override' pattern. The instructions do not themselves instruct exfiltration or reading unrelated system secrets; they focus on workspace files. Recommend confirming agents only run these checks with explicit user approval if agent autonomy is enabled.
Install Mechanism
No install spec; the skill is instruction-only with a small helper script. No remote downloads or package installs are present. Risk from installation is minimal.
Credentials
The skill declares no required environment variables, credentials, or config paths. References to config keys (e.g., memorySearch.extraPaths, OPENCLAW_PROFILE) are documentation only. The included script scans workspace markdown for the string 'sk-' (to detect likely API keys) and checks for .env — appropriate for an audit.
Persistence & Privilege
always is false and disable-model-invocation is default; the skill does not request permanent inclusion or modify other skills. The script is read-only in practice (checks and greps); it does not create or send data externally. No privileged actions or writes to other agent configs are present.
Scan Findings in Context
[system-prompt-override] expected: The scanner flagged text describing the platform's 'system prompt' used during automatic memory flush/compaction (present in references/openclaw-conventions.md). This is documentation of host behavior and appears expected for a workspace-conventions reference. However, any content that instructs an agent to honor or act on injected system prompts can be sensitive; treat such text as informational and verify the agent runtime enforces platform-level controls rather than skill-level prompts.
Assessment
This skill appears to be what it says: a local workspace auditor with a small helper script. Before installing/running: 1) Inspect the included scripts yourself (scripts/workspace-status.sh) and run them locally to see the output; they are simple read-only checks. 2) Verify there are no secrets in your workspace (MEMORY.md, memory/*.md, .env) — the script already greps for 'sk-' tokens and .env. 3) Be aware the docs describe automatic memory-flush behavior (agent writes to memory/YYYY-MM-DD.md during compaction); if you have concerns about automatic memory writes, review your agent's compaction/memoryFlush config. 4) If you enable autonomous invocation, restrict when the agent may run workspace audits (require explicit user confirmation) because the skill's instructions could be invoked repeatedly. Overall the package is coherent and low-risk, but always review included code before granting execution rights.Like a lobster shell, security has layers — review code before you run it.
latestvk973e4g7j7ysk0xwvfm3xyfsx180nmsf
License
MIT-0
Free to use, modify, and redistribute. No attribution required.