Risk & Remediation¶
Waymark's remediation layer sits above raw rollback. Instead of asking only “can this be undone?”, it scores risk, evaluates policy and compliance posture, and suggests safer alternatives when a full rollback is too risky.
Covered features¶
- F-22 — Risk assessment
- F-23 — Remediation recommender
- F-24 — Auto-block
- F-25 — Policy compliance evaluation
Risk score factors¶
Project documentation describes five major inputs to risk scoring:
- operation type
- scale (how many actions are involved)
- error pattern
- time since the actions ran
- system state / load
These combine into a score and a level such as low, medium, high, or critical.
Compliance evaluation¶
POST /api/remediation/evaluate-policy evaluates a session against built-in compliance frameworks documented in the repo:
- HIPAA
- SOC2
- PCI-DSS
- GDPR
The point is not to claim full certification from the tool alone. It is to encode review rules that help teams notice when a rollback or action sequence conflicts with their own control requirements.
Recommendation strategies¶
The remediation docs call out several strategies:
- partial rollback — undo only the safe subset
- staged rollback — execute in phases with checks between stages
- retry — for transient failures rather than destructive undo
- workaround — targeted fixes instead of full reversal
- escalation — route to an expert for judgment
Auto-block rules¶
High-risk sessions or policy violations can be blocked automatically. The documentation describes default thresholds such as risk scores above 7.0 and immediate blocking for disallowed policy conditions.
APIs¶
Assess a session¶
curl -X POST http://localhost:47000/api/remediation/assess \
-H "Content-Type: application/json" \
-d '{"session_id":"sess-123"}'
Evaluate policy and compliance¶
curl -X POST http://localhost:47000/api/remediation/evaluate-policy \
-H "Content-Type: application/json" \
-d '{"session_id":"sess-123"}'
Get ranked recommendations¶
curl -X POST http://localhost:47000/api/remediation/recommend \
-H "Content-Type: application/json" \
-d '{"session_id":"sess-123"}'
List active blocks¶
How this fits the rest of Waymark¶
Risk and remediation complement approval routing rather than replacing it. A team can:
- block obviously dangerous rollback attempts automatically
- require human approval on borderline cases
- offer safer alternatives than an all-or-nothing undo
- keep an audit trail of why a risky session was blocked or remediated
Remediation is about safer alternatives
The strongest value here is not just saying no. It is helping the team choose the least dangerous next step when a full rollback would create new problems.