Adversarial Test Plan —
Identification
- Target repo:
<org/repo@commit>
- Authorized scope source:
- Test window: to
- Test ID:
Threat model assumptions
- Trust boundaries:
- Adversary capabilities:
- In-scope assets:
- Out-of-scope assets:
- Known constraints (rate limits, fragile deps, etc.):
Selected checks
| Class |
Tool / model |
Rationale |
| SAST |
|
|
| DAST |
|
|
| Fuzzing |
|
|
| Dependency scan |
|
|
| Secret scan |
|
|
| Adversarial LLM (prompt) |
|
|
Runtime / environment profile
- Container image:
- CPU / memory limits:
- Network egress:
- Filesystem mounts:
Stop conditions
- High/critical finding requiring immediate halt:
- Resource limit exceeded:
- Suspected effect outside scope:
- Owner-requested abort:
Escalation triggers
- Real-world incident evidence outside test scope.
- Authentication failure suggesting unauthorized target.
- Sandbox escape suspicion.
Output expectations