Ashrafiucse Llm Security

Name: Ashrafiucse Llm Security
Author: ashrafiucse

Chains by @ashrafiucse v1.0.0 C PASS 1 flag

Audits LLM/AI application security — prompt injection and data exfiltration via tools, unsafe model/artifact deserialization (pickle, torch.load, LangChain unsafe_deserialization), hardcoded LLM API keys (OpenAI, Anthropic, Google, Groq, Hugging Face), over-powered agent tools (shell/REPL), prompt/PII logging, and telemetry capturing prompts (LangSmith, W&B). Use when auditing LangChain/LlamaIndex/AutoGen apps, OpenAI/Anthropic integrations, RAG pipelines, chatbots, or repos containing model files.

Install

$ cp -r cryptoskill/skills/chains/ashrafiucse-llm-security .claude/skills/

$ clawhub install ashrafiucse-llm-security

Quality Score

Score: 57/100 Grade: C Risk Gate: PASS

Trust profile trust grade not computed yet

1 red flag 9 cleared 1 not measured

Capabilities below are detected automatically by an open-source scanner that reads the skill's text and scripts (see how this is computed). Not measured means the scanner couldn't make a confident call — it is NOT a green check, and you should treat it as a possible red flag until a human or a stronger scanner has measured it.

1 Red flags things this skill can do that affect your security or funds

⚠
Can execute shell runs arbitrary shell commands on your machine

9 Cleared by scanner we scanned the skill's text & scripts and found no evidence of these

✓
Can move funds this skill can sign and send transactions on your behalf
✓
Requires private key you must hand over a private key, mnemonic, or wallet config
✓
Requires hosted operator depends on a third-party hosted service to function
✓
Uses remote install script setup pipes a remote shell script (curl | sh class)
✓
Mutable remote runtime runs remote code that can change behavior without a local diff
✓
Can install code installs software at setup time (npx, pip, brew, etc.)
✓
Can browse the web fetches arbitrary URLs at runtime
✓
Can write files writes to your local filesystem
✓
Can spawn sub-agents delegates to other skills or sub-agents

1 Not measured yet scanner couldn't make a confident call — treat as a possible red flag

?
Auto-invocable may be invoked by the agent without your explicit prompt

Execution mode

read_only Phase 1 single-mode classification — multi-mode breakdown deferred to Phase 2.

Ingredients (services this skill talks to)

We did not find any well-known hosted services in this skill's text or scripts. This does NOT mean the skill is local-only — it might use services we don't yet recognize, or talk to them through code paths our scanner can't reach. A complete dependency list (every package, library, and binary, with integrity hashes) is on the roadmap; today we only show recognized hosts.

Audits

No one has audited this skill yet. That is different from “audited and clean” — it just means no professional reviewer (a security firm, the CryptoSkill team, or a verified independent researcher) has signed off on it. There are no audit reports to read. How reviewer levels work →

SKILL.md → SOURCE.md → TRUST.auto.yaml → Browse directory →

Auto-generated by cryptoskill/extract-capabilities/0.3.1 · hosted-service list version 2026-09-06 · how this is computed

Source

View on GitHub →