# llm-security-hardening

> llm-security-hardening: 8 Keryn rules for python, javascript, typescript (1 error, 7 warning).

Source: https://keryn.dev/rules/llm-security-hardening/

8 rules · Keryn Software Licence · rules 0.8.0 · [all packs](https://keryn.dev/rules/)

**How to read a rule**

- **Severity**: How bad the finding is if it is real: `error` fails a default gate, `warning` and `info` do not.

- **Confidence**: **P0** exploitable as written; **P1** dangerous in context; **P2** a heuristic or a posture check: confirm by reading the code.

- **False-positive risk**: **low**: an exact API or flag, so hits are almost always real; **medium**: the pattern is legitimate in many programs, so check each hit; **high**: it finds a lead to confirm, so expect benign hits.

- **Edge**: The relationship the finding sits on. **reaches**: content or traffic gets to a thing; **calls**: one thing invokes another; **reads**: one thing reads from another; **writes**: one thing writes to another; **sends**: data leaves for another party; **grants**: authority is given or left unchecked; **depends_on**: something relies on a component it does not verify; **trusts**: one party believes another without evidence. The boundary it crosses, when there is one: **user->prompt** (untrusted input reaching a prompt); **model->code** (model output reaching code, a shell, a query or a page); **secret->outside** (a secret leaving the trusted side); **supply-chain** (code, models or prompts pulled in from outside).

- **Taxonomy chips**: References for readers who work in OWASP LLM 2025, OWASP Agentic, CWE or NIST AI RMF. They say where a rule belongs, not that a framework is satisfied.

## seceng.llm10.model-parameter-injection

`warning` python llm-security-hardening

Model parameters sourced from user input.

Confidence **P2**: a heuristic or a posture check: confirm by reading the code. False-positive risk **high**: it finds a lead to confirm, so expect benign hits.

Edge inspected: [reaches](https://keryn.dev/docs/edges/edge-kinds/#reaches) content or traffic gets to a thing; it crosses the [`user->prompt`](https://keryn.dev/docs/edges/trust-boundaries/#user-to-prompt) boundary (untrusted input reaching a prompt).

[LLM10](https://keryn.dev/taxonomies/owasp-llm/#llm10-unbounded-consumption) [ASI01](https://keryn.dev/taxonomies/owasp-agentic/#asi01-agent-goal-hijack) [ASI08](https://keryn.dev/taxonomies/owasp-agentic/#asi08-cascading-failures)

## seceng.llm07.persona-tampering

`error` python llm-security-hardening

Concatenation in system prompt.

Confidence **P1**: dangerous in context. False-positive risk **medium**: the pattern is legitimate in many programs, so check each hit.

Edge inspected: [trusts](https://keryn.dev/docs/edges/edge-kinds/#trusts) one party believes another without evidence.

[LLM07](https://keryn.dev/taxonomies/owasp-llm/#llm07-system-prompt-leakage) [ASI01](https://keryn.dev/taxonomies/owasp-agentic/#asi01-agent-goal-hijack)

## seceng.llm03.insecure-embedding-endpoint

`warning` python, javascript, typescript llm-security-hardening

Non-TLS embedding endpoint.

Confidence **P2**: a heuristic or a posture check: confirm by reading the code. False-positive risk **medium**: the pattern is legitimate in many programs, so check each hit.

Edge inspected: [grants](https://keryn.dev/docs/edges/edge-kinds/#grants) authority is given or left unchecked.

[LLM03](https://keryn.dev/taxonomies/owasp-llm/#llm03-supply-chain) [ASI06](https://keryn.dev/taxonomies/owasp-agentic/#asi06-memory-and-context-poisoning) [ASI04](https://keryn.dev/taxonomies/owasp-agentic/#asi04-agentic-supply-chain-vulnerabilities)

## seceng.llm10.model-routing-manipulation

`warning` python llm-security-hardening

The 'model' parameter is sourced from user input.

Confidence **P2**: a heuristic or a posture check: confirm by reading the code. False-positive risk **high**: it finds a lead to confirm, so expect benign hits.

Edge inspected: [reaches](https://keryn.dev/docs/edges/edge-kinds/#reaches) content or traffic gets to a thing; it crosses the [`user->prompt`](https://keryn.dev/docs/edges/trust-boundaries/#user-to-prompt) boundary (untrusted input reaching a prompt).

[LLM10](https://keryn.dev/taxonomies/owasp-llm/#llm10-unbounded-consumption) [ASI08](https://keryn.dev/taxonomies/owasp-agentic/#asi08-cascading-failures)

## seceng.llm10.model-routing-manipulation-js

`warning` javascript, typescript llm-security-hardening

The 'model' parameter is sourced from user input.

Confidence **P2**: a heuristic or a posture check: confirm by reading the code. False-positive risk **high**: it finds a lead to confirm, so expect benign hits.

Edge inspected: [reaches](https://keryn.dev/docs/edges/edge-kinds/#reaches) content or traffic gets to a thing; it crosses the [`user->prompt`](https://keryn.dev/docs/edges/trust-boundaries/#user-to-prompt) boundary (untrusted input reaching a prompt).

[LLM10](https://keryn.dev/taxonomies/owasp-llm/#llm10-unbounded-consumption) [ASI08](https://keryn.dev/taxonomies/owasp-agentic/#asi08-cascading-failures)

## seceng.llm07.insecure-prompt-storage

`warning` javascript, typescript llm-security-hardening

Potential system prompt or instructions being stored in insecure browser storage.

Confidence **P1**: dangerous in context. False-positive risk **medium**: the pattern is legitimate in many programs, so check each hit.

Edge inspected: [writes](https://keryn.dev/docs/edges/edge-kinds/#writes) one thing writes to another.

[LLM07](https://keryn.dev/taxonomies/owasp-llm/#llm07-system-prompt-leakage) [ASI01](https://keryn.dev/taxonomies/owasp-agentic/#asi01-agent-goal-hijack)

## seceng.llm02.openai-assistants-shared-thread

`warning` javascript, typescript, python openai

Hardcoded or constant thread\_id detected in OpenAI Assistants API.

Confidence **P2**: a heuristic or a posture check: confirm by reading the code. False-positive risk **high**: it finds a lead to confirm, so expect benign hits.

Edge inspected: [reads](https://keryn.dev/docs/edges/edge-kinds/#reads) one thing reads from another.

[LLM02](https://keryn.dev/taxonomies/owasp-llm/#llm02-sensitive-information-disclosure) [LLM07](https://keryn.dev/taxonomies/owasp-llm/#llm07-system-prompt-leakage) [ASI06](https://keryn.dev/taxonomies/owasp-agentic/#asi06-memory-and-context-poisoning)

## seceng.llm03.huggingface-token-hygiene

`warning` python huggingface

Hardcoded HuggingFace token detected.

Confidence **P2**: a heuristic or a posture check: confirm by reading the code. False-positive risk **medium**: the pattern is legitimate in many programs, so check each hit.

Edge inspected: [grants](https://keryn.dev/docs/edges/edge-kinds/#grants) authority is given or left unchecked.

[LLM03](https://keryn.dev/taxonomies/owasp-llm/#llm03-supply-chain) [ASI03](https://keryn.dev/taxonomies/owasp-agentic/#asi03-identity-and-privilege-abuse) [ASI04](https://keryn.dev/taxonomies/owasp-agentic/#asi04-agentic-supply-chain-vulnerabilities)
