Every identity gets a score per domain.
A high score means access in that domain is expected for this identity. Confidential or restricted access in a low-scoring domain that few peers hold is flagged for review. The score measures how well an identity's access fits its role.
Role and peers
Each department sets a starting expectation per domain. The score is 60% that expectation and 40% the share of peers who hold confidential or restricted access in the domain. Peers share a manager and title. When that group has fewer than three people, it widens toward the whole department.
Fit per grant
Each sensitive grant gets a fit: half the domain score, half the share of peers who reach that exact resource. Anomaly is sensitivity weight × (1 − fit), raised for access unused in 90 days, machines without an owner, and contractors on restricted systems. A grant is flagged at 0.45 and marked high at 0.70.
History
Job history shows where the person worked when each grant was made. Access granted while they were in another department counts for more, because many flags are leftovers from a job the person no longer has.
Clearance Score, in practice.
- A score from 0 to 100 per domain for every person, service account, and AI agent
- Flags weighted by sensitivity: restricted counts fully, confidential 0.6, internal 0.2, and public access is never flagged
- No flag when half or more of a person's peers hold the same access
- Approvals that hold for one year, with an optional reason
Check access before you grant it.
The pre-grant check will hook into access requests in Slack or your identity provider. Before an approver clicks, they will see the score, the peers, and whether the grant would be flagged.
Request arrives
Requests will come from Slack, from your identity provider's access request flow, or from a ticket. Each one will be matched to an identity and a resource on the graph.
Commando scores it
The scoring that runs on existing grants will run on the request as if it were already granted. The approver will see the verdict, the reasons, and any path that already reaches the resource.
Approver decides with evidence
The approver will grant or deny in the system they already use. 1Commando only reads from connected systems, so a new grant shows up on the graph after the next import.
- A verdict before access is granted
- A warning when the person already reaches the resource through a group
- The same scores, peers, and reasons reviewers see in a review
- A record of each request and its verdict
One number for the whole company.
The posture score rolls the fit of every sensitive grant into one figure for the company and one for each domain. The overview shows the company figure and the six domains with the most open flags, and each of those domains links to the flags behind it.
Roll up
Posture is 100 × (1 − Σ weight × anomaly ÷ Σ weight) over every sensitive grant held by a current identity. The same sum per domain gives each domain its score.
Weighting
Restricted access has weight 1, confidential 0.6, and internal 0.2, so restricted access counts five times as much as internal access.
Decisions
Approved grants and exceptions count as fitting until they expire. A removal request keeps counting until the next import shows the access is gone at the source.
- A company score from 0 to 100
- A score for each domain, such as Finance or Engineering
- The flags behind each listed domain, one click from the overview
- Credit for a removal only once the access is gone at the source
