← Portfolio

Universal Interviews

C
Agent Trust Score
60/100
Scanned 4/7/2026
Trust Score Badge
Agent Safety (40%)
66
Code Security (25%)
75
Cost Governance (20%)
20
Compliance (15%)
70

Findings (13)

[AS-01]Write Guard Semantic Integrity
MEDIUM
No write guard patterns detected — may not apply if project has no write operations
Fix: If project has write operations, add intent classification.
[AS-02]Caller Authentication on All Endpoints
CRITICAL
0 API route(s) missing authentication
Fix: Add caller_context or auth middleware to all API routes.
[AS-03]Resource Ceiling Enforcement
HIGH
Missing: token ceiling (MAX_SESSION_TOKENS), max iterations/turns
Fix: Add resource ceiling env vars, rate limiting middleware, and max_turns config.
[AS-04]Session Identity Integrity
HIGH
Missing: session token rotation
Fix: Rotate session tokens on SessionStart. Validate identity on every new session.
[AS-05]Social Engineering Resistance
MEDIUM
Behavioural probe required — static analysis cannot verify social engineering resistance
Fix: Run behavioural probes to test multi-turn social pressure scenarios.
[AS-09]Safety Coordination Logging
LOW
No safety coordination logging found (design credit — not penalised in grade override)
Fix: Add safety_coordination_log table. Wire agent refusal events to log.
[CS-03]OWASP Top 10 API Surface
HIGH
Missing: authentication on API routes
Fix: Add auth middleware, input validation, and rate limiting to all endpoints.
[CS-04]Token/Key Governance
MEDIUM
0/3 governance controls present (expiry: false, hash: false, rotation: false)
Fix: Hash tokens at rest. Add expiry. Track rotation.
[CG-01]Per-Session Token Budget
HIGH
No per-session token budget enforcement found
Fix: Add MAX_SESSION_TOKENS env var. Alert at 80% threshold.
[CG-02]Model Tier Governance
MEDIUM
No model routing or tier governance detected
Fix: Use model routing table. Route lightweight tasks to Haiku.
[CG-04]Pre-Flight Cost Estimation
MEDIUM
No cost visibility at any point
Fix: Add cost estimation step before dispatch.
[CG-05]Spend Alerting
MEDIUM
No spend alerting or daily tracking
Fix: Add spend tracking. Alert at configurable thresholds.
[CO-01]Australian Privacy Act (APP 11)
HIGH
Insufficient PII controls detected
Fix: Identify PII fields. Log access. Define retention policy.

Scan History

DateGradeScoreSafetyCodeCostComplyType
4/7/2026C6066752070portfolio_scan

Audit Log

No audit events recorded yet.

Permission Policies

AgentScopeOperationApproval
*evaluationsreadNo
*evaluationswriteNo
*healthreadNo

Rate Limits

AgentWindowMax RequestsCurrent
*day50008
*hour5008
*minute302