Skills/ClawHub/improvement-discriminator

improvement-discriminator@1.1.1

CAUTION· static only

Publisher claims: "当需要对改进候选多人盲审打分、用 LLM 做语义评估、判断候选是否应被接受、或打分结果全是 hold 想知道为什么时使用。支持 --panel 多审阅者盲审和 --llm-judge 语义评估。不用于结构评估(用 improvement-learner)或门禁决策(用 improvement-gate)。" Observed in sandbox: static analysis only — behavior not yet observed.

Registry ClawHubPublisher lanyasheng · publishing for 227 daysInstalls 891Last detonated 2 h ago

Coverage

Static only; nothing was run yet.

The label rests on the code, the instruction files and the publisher. The honeypot run replaces this verdict when it completes.

Is [email protected] safe to install?

KillerSkills analysed it statically (a sandbox run is pending) on 2026-11-19 and scored it 68/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Summary: static analysis only — behavior not yet observed.

Behavior trace

Queued for honeypot run · static analysis only
t+CategoryObserved action

This version has not been detonated yet. The score above counts behavior as unknown (50/100), so it cannot read Verified until the honeypot run completes and replaces this verdict.

Static findings

MEDIUM
File type combined with upload actionSKILL.md:105 · File type combined with upload action — agent may exfiltrate local data files--input candidates.json --panel --llm-judge mock --output scored.jsonFix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
MEDIUM
Python exec or eval of computed codescripts/score.py:402 · Python exec or eval of computed coderisk_keywords = ["password", "token", "secret", "api_key", "credential", "auth", "exec(", "eval(", "os.system"]Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST05
LOW
Dynamic module loadinginterfaces/critic_engine.py:377 · Dynamic module loadingself.skill_module = importlib.util.module_from_spec(spec)Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
File type combined with upload actionSKILL.md:77 · File type combined with upload action — agent may exfiltrate local data files$ python3 scripts/score.py --input candidates.json --panel --llm-judge mock --output scored.jsonFix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
LOW
Dynamic module loadinginterfaces/critic_engine.py:379 · Dynamic module loading (importlib.loader.exec_module via provenance)spec.loader.exec_module(self.skill_module)Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
File type combined with upload actionSKILL.md:101 · File type combined with upload action — agent may exfiltrate local data filespython3 scripts/score.py --input candidates.json --output scored.jsonFix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07

Cost to the agent

Instructions loaded
≈1,481 tok
1 file · 15 in bundle

Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.

Known vulnerabilities

Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.

What would make this Verified

12 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%
  1. Unclaimed publisher≈ +8 on the totalClaim lanyasheng: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
  2. File type combined with upload action≈ +2 on the totalRemove the call, or declare the host in frontmatter and explain in the README what is sent and why.
  3. Python exec or eval of computed code≈ +2 on the totalPut hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
  4. Dynamic module loading≈ +1 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
Publish a new version to ClawHub, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim lanyashengPropose a patch to the maintainer