claw-lark Patches

PassAudited by ClawScan on May 1, 2026.

Overview

The skill is transparent about patching claw-lark plugin files, but users should only run it intentionally because it changes installed OpenClaw extension behavior.

This skill appears coherent and purpose-aligned, but it directly edits installed claw-lark plugin files. Before running it, review the patch script, make sure the target path is correct, set BOT_OPEN_ID and BOT_NAME carefully if needed, and be prepared to reinstall or restore claw-lark if the patched behavior is not what you want.

Findings (3)

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.

What this means

Running the script changes installed claw-lark runtime files, so the gateway may behave differently after restart.

Why it was flagged

The quick-apply script executes embedded Node.js code and writes patched JavaScript back into the claw-lark plugin files. This is disclosed and central to the skill's purpose, but it is still local code-modification authority.

Skill content
node << 'PATCH_EOF' ... fs.writeFileSync(path, content);
Recommendation

Run the script only when you intend to patch claw-lark, review the script first, and keep a way to restore the plugin by reinstalling or backing up the original files.

What this means

After patching, outbound Lark sends, reactions, and media operations may use the default configured Lark account where they previously failed.

Why it was flagged

The documented patch changes outbound methods to resolve and use the configured default Lark account when an account is not passed directly. This is purpose-aligned for fixing claw-lark messaging, but it affects which delegated Lark account is used.

Skill content
const account = args.account ?? resolveAccount(args.cfg, args.accountId ?? "default");
Recommendation

Confirm the configured claw-lark default account has the intended workspace permissions and that only trusted users or workflows can trigger outbound Lark actions.

What this means

Users have less external provenance information and may not realize the quick-apply script depends on Node being available.

Why it was flagged

The registry metadata does not provide provenance or declare the Node runtime used by the included patch script. This is not evidence of malicious behavior, but it is useful context for a script that modifies installed plugin code.

Skill content
Source: unknown; Homepage: none; Required binaries (all must exist): none; No install spec — this is an instruction-only skill.
Recommendation

Inspect the included script before running it and verify it comes from a source you trust.