Illustrative engineering portfolio
See where agent performance has been reported, where it remains unknown, and which codebase constraints need attention. Open a repository for task-level evidence and the supporting engineering baseline.
Proof before rollout
No estate-wide success average: repositories have different tasks and configurations. Illustrative, stale and local-command data do not count as verified performance.
| Repository | Evidence status | Tasks / runs | Verified completion |
|---|---|---|---|
| payments-core | not measured | 0 / 0 | Not established |
| customer-platform | not measured | 0 / 0 | Not established |
| policy-api | not measured | 0 / 0 | Not established |
| docs-site | not measured | 0 / 0 | Not established |
Where attention goes first
Priority combines criticality, static-score gap, production status, and critical caps. This index helps order remediation priorities; it does not predict outages or rank team performance.
| Repository / system | Business context | Static score | Evidence / execution | Next decision | Priority index |
|---|---|---|---|---|---|
| payments-corePayments platform · Payments team | critical · production service · high | 30/100 · L1 | 100% stable-check coverage · execution not run | foundation first cap: SAFE-01 | 400 |
| customer-platformCustomer experience · Experience team | high · production service | 84/100 · L2 | 100% stable-check coverage · execution not run | task-trial candidate | 68 |
| policy-apiPolicy services · Core platform | high · production service | 100/100 · L4 | 100% stable-check coverage · execution not run | task-trial candidate | 20 |
| docs-siteDeveloper documentation · Developer experience | low documentation | 24/100 · L0 | 100% stable-check coverage · execution not run | profile review | Not ranked |
Not ranked under the current code-centric rubric: docs-site (low, documentation). Review applicability before using their scores for decisions.
Systemic fixes
- ENV-03 appears in 2 repositories: customer-platform, payments-core
- ENV-04 appears in 2 repositories: customer-platform, payments-core
- GIT-01 appears in 2 repositories: customer-platform, payments-core
- SAFE-02 appears in 2 repositories: customer-platform, payments-core
Trials stay separate
No empirical agent-task trials were supplied. Candidate zones do not establish task success.
What needs an owner
- critical · payments-core: Static score capped by SAFE-01. Review the cited finding and remediate before a task trial.
- high · payments-core: Production system with weak static controls. Validate release and rollback controls, then address top repository findings.
90-day improvement plan
Days 0–30 · Close critical risks and evidence gaps
- payments-core: Review the cited finding and remediate before a task trial.
- payments-core: Validate release and rollback controls, then address top repository findings.
Days 31–60 · Remove recurring constraints
- Address ENV-03 across 2 repositories, then rerun the same standard.
- Address ENV-04 across 2 repositories, then rerun the same standard.
- Address GIT-01 across 2 repositories, then rerun the same standard.
- Address SAFE-02 across 2 repositories, then rerun the same standard.
Days 61–90 · Prove movement and bound agent use
- Reassess with the same rubric and separate rubric effects from repository effects.
- Run named, bounded agent-task trials only for agreed candidates.
- Review deployment and operational evidence before any production-readiness claim.
A system owner should validate effort, dependencies, and sequencing before committing dates.
Priority index: criticality weight (critical 4, high 3, medium 2, low 1) × (100 − static score), plus 20 for production and 100 for a cap. Documentation, archived, monorepo, and unknown profiles are not ranked until their rubric applicability is reviewed. No cross-repository average is reported. Repository profiles may be manually corrected in the manifest. Static structural diagnostics do not change the score. Execution, when separately requested, is not a security sandbox.