Skillfile pilot offer
Audit AI-agent access before rollout.
For security and platform teams reviewing AI-agent access before production: Skillfile starts with one bounded audit of coding agents, MCP servers, deploy paths, and customer-data-adjacent systems.
What the pilot gives you
A practical path from unmanaged agent configs to reviewable capability boundaries, without waiting for a full platform rollout.
Current access map
Inventory visible agent configs, MCP endpoints, local rules, workflow files, and tool grants so the review starts from evidence.
Risk register
Separate allowed, human-gated, and blocked capabilities. Flag deploy paths, network egress, secret-adjacent access, and customer-data exposure.
Governed .skill drafts
Turn the first approved workflows into reviewable manifests with permissions, integrations, and audit-ready rationale.
Rollout decision
Leave with a clear yes/no decision: keep the team pilot running, expand to more teams, or use the audit packet internally.
Pilot pricing
Start with the Access Audit Sprint. The monthly pilot begins after that audit only when there is a real rollout path.
Access Audit Sprint
$2,500
Fixed-scope starting point
- One repo or agent profile
- Config-only scan and unmanaged access map
- Risk register with allowed/gated/blocked recommendations
- Three governed
.skilldrafts
Team Governance Pilot
$1,500/mo
Begins after Access Audit Sprint if rollout is real
- One team workspace
- Approval queue for risky grants
- Sync guidance for agent surfaces
- Audit evidence exports
Company Rollout
from $5,000/mo
For multi-team governance
- Multiple teams and policy packs
- Custom integrations and rollout support
- Executive/control-plane reporting
- Security and platform stakeholder reviews
How the work runs
The same loop the product demonstrates: scan, expose, approve, sync, and audit. Human review stays in the path.
- ScanBring one repo, agent config, or MCP profile. We inspect configuration and visible policy files, not secrets.
- ExposeWe show unmanaged access, risky grants, blocked patterns, and the review questions each owner needs to answer.
- ApproveApproved behavior becomes a `.skill` capability boundary with explicit permissions and rationale.
- SyncThe approved path is prepared for your target agent surfaces instead of living as scattered prompt/config fragments.
- AuditYou get the evidence packet: what changed, who should own it, what stayed blocked, and what to do next.
Access audit request
Bring one repo or agent profile.
We will send next steps, confirm scope, and keep the first audit bounded. If the audit shows a real rollout path, the monthly pilot can start after it. No newsletter or sales drip.