Skip to content
ai-supply.store
탐색카테고리리더보드커뮤니티Agent APIFAQ
로그인무료 가입
catalog / Coding / LLM (CLI)
⇄ConnectorCodingFree

LLM (CLI)

CLI and Python library to access many LLM providers with a plugin system and SQLite logging.

@ai-supply
설치 수47k
↗ 소스 저장소
← More CodingCoding leaderboard →How we grade security →Source ↗

LLM (simonw/llm)

LLM is a command-line tool and Python library by Simon Willison for interacting with large language models — from OpenAI and Anthropic to dozens of local and hosted models via a plugin system.

It supports prompting, chat, embeddings, and logging every interaction to a local SQLite database, and its plugin architecture lets you add new model providers and tools.

Apache-2.0 licensed and lightweight; a versatile connector for scripting and piping LLMs into shell workflows.

Rating rank
#1
of 27 in Coding
Install rank
#12
of 27 in Coding
Security score
100/100 · A
safe
Security rank
#1
of 27 in Coding
Installs
47k
cat avg 157k
This listing vs category average
Installs
this
cat avg
Security (of 100)
this
cat avg
Adoption trend
See the Coding leaderboard →
✓ Security: Safe · 100100/100 · grade Ascanned 15h ago
✓ no compromise signals21 risk-surface · 8/20 OWASP controls flagged

Compromise signals — malicious or tampered code (leaked secrets, backdoors, a dropped executable) — reduce the score, and known dependency CVEs carry a bounded penalty (they warrant review but never QUARANTINE — update the dependency to clear). Other dangerous-by-capability traits are risk surface, expected for some capabilities. Every finding is mapped to its OWASP control below.

What this capability can do · high confidence (static)
Tools (6)
uppercount_charsllm_versionoutput_as_jsonllm_timeno_impl
⚑ filesystem⚑ shell⚑ network⚑ secrets
egress → img.shields.io, pypi.org, llm.datasette.io, datasette.io, formulae.brew.sh, www.youtube.com, simonwillison.net, pypa.github.io +32
auth: api_keyimg.shields.iogithub.compypi.orgllm.datasette.iodatasette.ioformulae.brew.shwww.youtube.comsimonwillison.net

Findings mapped to the OWASP Top 10 for LLM Applications (2025) and the OWASP Machine Learning Security Top 10. Expand any flagged control for the exact findings — compromise reduces the score; expected/risk-surface do not.

OWASP Top 10 for LLM Applications
⚠LLM01Prompt Injectionhigh
Adversarial instructions embedded in an artifact that hijack a downstream LLM.
•Prompt-injection phrasing — instruction-subversion language detected · simonw-llm-dfe1278/docs/tools.md (CWE-77)risk surface
⚠LLM06Excessive Agencyhigh
Over-broad tool/permission surface or unrestricted egress.
•External endpoints declared — 1 distinct host(s) · simonw-llm-dfe1278/LICENSErisk surface
•Broad capability surface — 3 high-impact capability categories referenced — verify least-privilege · simonw-llm-dfe1278/README.md (CWE-272)risk surface
•External endpoints declared — 14 distinct host(s) · simonw-llm-dfe1278/README.mdrisk surface
•Broad capability surface — 4 high-impact capability categories referenced — verify least-privilege · simonw-llm-dfe1278/docs/changelog.md (CWE-272)risk surface
•External endpoints declared — 21 distinct host(s) · simonw-llm-dfe1278/docs/changelog.mdrisk surface
•External endpoints declared — 3 distinct host(s) · simonw-llm-dfe1278/docs/contributing.mdrisk surface
•External endpoints declared — 2 distinct host(s) · simonw-llm-dfe1278/docs/embeddings/index.mdrisk surface
•External endpoints declared — 4 distinct host(s) · simonw-llm-dfe1278/docs/embeddings/writing-plugins.mdrisk surface
•External endpoints declared — 6 distinct host(s) · simonw-llm-dfe1278/docs/other-models.mdrisk surface
•External endpoints declared — 35 distinct host(s) · simonw-llm-dfe1278/docs/plugins/directory.mdrisk surface
•External endpoints declared — 10 distinct host(s) · simonw-llm-dfe1278/docs/plugins/tutorial-model-plugin.mdrisk surface
•Egress to a private/loopback host — 127.0.0.1 · simonw-llm-dfe1278/docs/schemas.md (CWE-918)risk surface
•External endpoints declared — 5 distinct host(s) · simonw-llm-dfe1278/docs/templates.mdrisk surface
⚠LLM05Improper Output Handlingmedium
Code that pipes model/user output into shell, eval, SQL or paths unsafely.
•Suspicious code patterns — OS command execution · simonw-llm-dfe1278/README.md (CWE-78)risk surface
•Path traversal sequences — '../' present in content or name · simonw-llm-dfe1278/docs/plugins/tutorial-model-plugin.md (CWE-22)risk surface
•Suspicious code patterns — dynamic code execution · simonw-llm-dfe1278/llm/cli.py (CWE-95)risk surface
⚠LLM10Unbounded Consumptionmedium
Unbounded loops/recursion causing DoS or runaway cost.
Enforced at runtime by the gateway (rate limits + spend caps + size caps); static check flags unbounded loops.
•Potentially unbounded loop — an infinite loop (while True / while(1) / for(;;)) may cause runaway consumption · simonw-llm-dfe1278/llm/cli.py (CWE-835)risk surface
⚠LLM03Supply Chainlow
Vulnerable/compromised dependencies, models or archives in the artifact.
•Dependency manifest — 7 pip requirements declared · simonw-llm-dfe1278/docs/requirements.txtrisk surface
§LLM09MisinformationGovernance
Artifacts designed to produce false/deceptive output.
Detectable only by runtime behavioral evaluation; addressed via responsible-use attestation.
✓LLM02Sensitive Information DisclosurePassed
✓LLM04Data and Model PoisoningPassed
Backdoors/poisoning in training data or serialized models.
Behavioral poisoning needs model execution; static check covers unsafe serialization + dataset skew only.
✓LLM07System Prompt LeakagePassed
✓LLM08Vector and Embedding WeaknessesPassed
PII or plaintext source leakage in embedding/vector exports.
Embedding inversion/poisoning is largely runtime; static check covers PII in vector exports.
OWASP Machine Learning Security Top 10
⚠ML02Data Poisoninghigh
Poisoned training datasets with triggers or anomalous distributions.
Static check covers trigger phrasing, PII and label skew; full poisoning detection is runtime.
•Prompt-injection phrasing — instruction-subversion language detected · simonw-llm-dfe1278/docs/tools.md (CWE-77)risk surface
⚠ML09Output Integritymedium
Middleware tampering with model outputs in transit.
Gateway enforces TLS + response integrity; static check flags output-rewriting code.
•Suspicious code patterns — OS command execution · simonw-llm-dfe1278/README.md (CWE-78)risk surface
•Path traversal sequences — '../' present in content or name · simonw-llm-dfe1278/docs/plugins/tutorial-model-plugin.md (CWE-22)risk surface
•Suspicious code patterns — dynamic code execution · simonw-llm-dfe1278/llm/cli.py (CWE-95)risk surface
⚠ML06AI Supply Chainlow
Compromised PyPI/npm packages, typosquats, unsafe serialized models.
•Dependency manifest — 7 pip requirements declared · simonw-llm-dfe1278/docs/requirements.txtrisk surface
§ML01Input Manipulation (Adversarial)Governance
Models vulnerable to adversarial perturbations.
Requires runtime robustness evaluation; addressed via publisher robustness attestation.
§ML03Model InversionGovernance
Training data reconstructable from a model's outputs.
Runtime/evaluation property; addressed via model-card data-provenance + DP attestation.
§ML04Membership InferenceGovernance
Determining whether a record was in the training set.
Runtime/evaluation property; addressed via overfitting disclosure + DP attestation.
§ML08Model SkewingGovernance
Models trained on skewed data producing biased output.
Requires fairness evaluation; addressed via model-card bias/limitations disclosure.
✓ML05Model TheftPassed
Unlicensed re-distribution / license-incompatible derivatives.
Static check verifies license declaration; extraction throttling is runtime.
✓ML07Transfer Learning AttackPassed
Backdoored base models / LoRA adapters propagating to derivatives.
Backdoor detection needs behavioral probing; static check covers unsafe serialization + provenance.
✓ML10Model Poisoning (Weights)Passed
Tampered model weight files; integrity must be verifiable.
Static check enforces safe formats + records a content hash for downstream verification.
Other findings (7) · hygiene / uncategorized
•Unrecognized file type — '.gitignore' is not on the allowlist · simonw-llm-dfe1278/.gitignorerisk surface
•Unrecognized file type — '.?' is not on the allowlist · simonw-llm-dfe1278/Justfilerisk surface
•Unrecognized file type — '.in' is not on the allowlist · simonw-llm-dfe1278/MANIFEST.inrisk surface
•Suspicious network references — suspicious TLD (15 URLs) · simonw-llm-dfe1278/docs/plugins/tutorial-model-plugin.mdrisk surface
•Suspicious network references — raw IP URL (16 URLs) · simonw-llm-dfe1278/docs/schemas.mdrisk surface
•Unrecognized file type — '.ini' is not on the allowlist · simonw-llm-dfe1278/mypy.inirisk surface
•Possible obfuscation — very long lines · simonw-llm-dfe1278/tests/cassettes/test_openai_responses/test_responses_interleaved_reasoning_between_tool_calls.yamlrisk surface
✔ verified source · pinned simonw-llm-dfe1278
Check against a policy

The same gate an agent runs before installing (POST /api/v1/trust/llm-cli/check). Click a policy:

Consume LLM (CLI) programmatically. Authenticate with an API key or session — see Authorize an agent.

# Agents: CHECK BEFORE YOU INSTALL (no auth) — score, grade, level, capability manifest
curl https://ai-supply.store/api/v1/trust/llm-cli

# Gate against your org policy (returns { pass, violations })
curl -X POST https://ai-supply.store/api/v1/trust/llm-cli/check \
  -H "Content-Type: application/json" \
  -d '{"minGrade":"B","denyPermissions":["shell"],"denyUnknownEgress":true}'

# CLI
npx ai-supply add llm-cli

# REST (install → download)
curl -X POST https://ai-supply.store/api/v1/listings/llm-cli/install \
  -H "Authorization: Bearer $AIM_KEY"

# MCP tool
install_listing({ "slug": "llm-cli" })
OpenAPI spec →
vlatest
✓ Security: Safe · 10017h ago

Curated mirror — latest upstream source. See the repository for tagged releases.

Sign in and install this listing to leave a review.

More from @ai-supply

View profile →
◉Agent
MetaGPT
Multi-agent framework that assigns GPT roles (PM, engineer, QA) to solve complex software tasks end-to-end.
↓ 1.0M
⇄Connector
vLLM
High-throughput, memory-efficient LLM inference engine with PagedAttention and continuous batching.
↓ 892k
⇄Connector
Meilisearch
Lightning-fast open-source search engine with typo-tolerance, semantic hybrid search, and sub-50ms response times.
↓ 811k
△Eval
Weights & Biases (wandb)
ML experiment tracking and visualization — log metrics, hyperparameters, models, and media in real time.
↓ 784k
ai-supply.store

무료로 제공하는 보안 검증 AI 역량 — skill, MCP, plugin, agent, 데이터셋을 비롯한 모든 항목에 보안 점수를 매기고 최신성을 추적하며, 사람과 agent 모두를 위해 만들었습니다.

api · v3.1status · all green
문의하기
support@ai-supply.storesecurity@ai-supply.store
카탈로그
  • 탐색
  • 카테고리
  • 리더보드
  • 벤치마크
  • 보안
커뮤니티
  • 커뮤니티
  • FAQ
에이전트용
  • 빠른 시작 (60s)
  • 에이전트 승인
  • Agent API
  • OpenAPI 사양
빌더용
  • 게시
  • 대시보드
계정
  • 계정 만들기
  • 로그인
  • 설정
법적 정보
  • 이용약관
  • 게시자 계약
  • 이용 정책
  • 개인정보 처리방침