---
id: owasp-asi01-agent-goal-hijack
track: coding
version: 0.0.2
label: advisory
keyword: advise
where: text
sha256: 4dd9952dcd8927d9691e2498a5757ac6699aefd91ada629ec3e3f6e59d927cee
source: https://github.com/open-coder-ai/chock-catalog/tree/1e60d211c1bb767152c6ab0167932da06ab3a10f/agentic-security/owasp-asi01-agent-goal-hijack
catalog_ref: 1e60d211c1bb767152c6ab0167932da06ab3a10f
---

# OWASP ASI01 — Agent Goal Hijack

owasp-asi01-agent-goal-hijack · version 0.0.2

Advisory — Advisory only. It is skill text, and nothing stops a violation. Chock adds friction. It is not a security boundary. 

In a git repo: advisory (The agent reads the text, and nothing stops a violation.)



## What it does

Keep an agent's objective under the operator's control when the agent ingests untrusted content. Separate retrieved data from instructions, refuse tool-scope expansion requested by that data, and confirm sensitive actions against the raw action rather than a summary. Use when building RAG pipelines, email/ticket/doc readers, browser agents, or any planner whose context includes fetched content. Do NOT use for the coding agent's own session hygiene — that is \`injection-defense\`.

## Its label on each agent

- Claude Code (tested): Advisory — Advisory only. It is skill text, and nothing stops a violation. Builder: /coding/builder/#client=claude-code&p=owasp-asi01-agent-goal-hijack
- Cursor (tested): Advisory — Advisory only. It is skill text, and nothing stops a violation. Builder: /coding/builder/#client=cursor&p=owasp-asi01-agent-goal-hijack
- Codex (tested): Advisory — Advisory only. It is skill text, and nothing stops a violation. Builder: /coding/builder/#client=codex&p=owasp-asi01-agent-goal-hijack
- GitHub Copilot (tested): Advisory — Advisory only. It is skill text, and nothing stops a violation. Builder: /coding/builder/#client=copilot&p=owasp-asi01-agent-goal-hijack
- Devin (tested): Advisory — Advisory only. It is skill text, and nothing stops a violation. Builder: /coding/builder/#client=devin&p=owasp-asi01-agent-goal-hijack
- Grok (tested): Not graded yet for Grok. Builder: /coding/builder/#client=grok&p=owasp-asi01-agent-goal-hijack
- goose (tested): Not graded yet for goose. Builder: /coding/builder/#client=goose&p=owasp-asi01-agent-goal-hijack
- Antigravity (tested): Not graded yet for Antigravity. Builder: /coding/builder/#client=antigravity&p=owasp-asi01-agent-goal-hijack
- OpenHands (untested): Not graded yet for OpenHands. Builder: /coding/builder/#client=openhands&p=owasp-asi01-agent-goal-hijack
- Kimi Code (untested): Not graded yet for Kimi Code. Builder: /coding/builder/#client=kimi-code&p=owasp-asi01-agent-goal-hijack
- Factory Droid (untested): Not graded yet for Factory Droid. Builder: /coding/builder/#client=factory-droid&p=owasp-asi01-agent-goal-hijack

## What it misses

This policy does not state what it misses.

## Latest change

0.0.2 · 2026-08-16 — Add compliance claims (mitre_atlas); no mechanism change.

## Provenance

- Tested by 0 of 6 cases run automatically
- SHA-256: `4dd9952dcd8927d9691e2498a5757ac6699aefd91ada629ec3e3f6e59d927cee`
- Catalog commit: https://github.com/open-coder-ai/chock-catalog/commit/1e60d211c1bb767152c6ab0167932da06ab3a10f
- Source: https://github.com/open-coder-ai/chock-catalog/tree/1e60d211c1bb767152c6ab0167932da06ab3a10f/agentic-security/owasp-asi01-agent-goal-hijack (catalog path `agentic-security/owasp-asi01-agent-goal-hijack`)
- Presets: OWASP agentic top 10
- JSON: `/api/coding/policies/owasp-asi01-agent-goal-hijack.json`
- Report a problem: https://chock.sh/report/?policy=owasp-asi01-agent-goal-hijack&version=0.0.2&page=%2Fcoding%2Fpolicies%2Fowasp-asi01-agent-goal-hijack%2F