1.0 Scope

This app is for Bethesda mod analysis, recommendations, and dev support. High-risk automation is behind explicit tier gates and guardrails.

2.0 Constraint model (organized disclosure)

2.1 Authorization constraints

• Login + tier gate + OpenClaw enabled flag.
• Short-lived grant token required for OpenClaw actions.
• Optional same-IP binding for grant verification.

2.2 Path and file constraints

• Only safe relative paths inside dedicated sandbox workspace.
• Path length, depth, extension allowlist, and deny-segment checks enforced.
• Hard block for system-critical path segments and shell-like path tokens.

2.3 Operation constraints

• Allowlisted operations only.
• Bytes and file-touch limits enforced.
• Dry-run guard-check endpoint available before execution.

2.4 Audit constraints

• OpenClaw events are logged for safety forensics.
• Grant verification failures and blocked actions are recorded.

3.0 Information processing contract

Data enters via mod lists, dev artifacts, optional telemetry, and community activity. Processing uses parser/ranker/AI layers. Action outputs are constrained by role/tier/safety policy.

See machine-readable flow map: /api/information-map.

4.0 Future goal disclosure

Live modding-as-you-play is an eventual R&D direction as AI and local safety tooling improve. Current implementation focuses on guarded planning loops, explicit permission scopes, and preflight validation before action.

5.0 Recommended user workflow

1. Run analyze and fix high-severity issues first.
2. Generate plan proposal and review every step.
3. Run guard-check and confirm backups.
4. Execute only sandbox-safe actions.
5. Submit loop feedback and iterate one controlled change at a time.

Terms · Privacy · API Hub