Skip to content

Dashboard Pages ​

Detailed description of each dashboard page.

Overview (/) ​

Fleet-wide metrics:

  • Total repos, issues, PRs, CI runs
  • CI success rate (pie chart)
  • Recent activity feed
  • Installation health indicators

Repos (/repos) ​

Sortable, filterable table of all repositories:

  • Repo name, owner, language, visibility
  • Open issues/PRs count
  • Last synced timestamp
  • Sync trigger button

Issues (/issues) ​

Issue list with triage columns:

  • Title, number, repo, state
  • Triage type (bug/feature/etc.)
  • Priority (critical/high/medium/low)
  • Labels
  • Pagination

Pull Requests (/pull-requests) ​

PR list with size/risk:

  • Title, number, repo, state
  • Head/base branch
  • Triage size, risk rating
  • Draft status

CI (/ci) ​

Two tabs:

  • CI Runs — All workflow runs with conclusion, healing status
  • Heal History — Auto-generated patch PRs with success/fail stats

Insights (/insights) ​

Charts and metrics:

  • Issue/PR velocity over time (line chart)
  • CI pass/fail trends (area chart)
  • Repository health scores
  • Merge time distribution

Maintainer (/maintainer) ​

Five tabs:

  • Settings — Per-repo maintainer config
  • Stale — Stale items with warn/close actions
  • Branches — Branch rules and cleanup
  • Collaborators — Repo access management
  • Audit Log — Governance actions

Fix Attempts (/fix-attempts) ​

Autonomous contributor history:

  • Issue number, repo, status
  • Complexity, confidence
  • PR link (if created)
  • Error details (if failed)

Duplicates (/duplicates) ​

Duplicate detection dashboard:

  • Pending/confirmed/dismissed signals
  • Similarity scores
  • Confirm/dismiss actions
  • Embedding coverage stats

Automation (/automation) ​

Four tabs:

  • Merge Queue — Queue entries and config
  • Feedback — Notification rules
  • Telemetry — Pipeline metrics and events
  • Rollbacks — Rollback history

Trust (/trust) ​

Three tabs:

  • Flaky Tests — Detected flaky tests with quarantine actions
  • Dependencies — Manifests and vulnerability advisories
  • Policies — Reconciliation runs and drift status

Intelligence (/intelligence) ​

Two tabs:

  • AI Review — Review results, findings, config
  • Audit Trail — Chain-verified entries, compliance reports

Config (/config) ​

Per-repository configuration management:

  • Repo selector — Dropdown to pick from synced repos
  • Pillar cards (9) — Toggle each pillar on/off, configure sub-options:
    • Issue & PR Triage: auto_label, auto_comment, duplicate_detection, triggers
    • Self-Healing CI: auto_patch, max_fix_attempts, min_confidence_to_patch, triggers
    • Autonomous Contributor: max_file_changes, max_line_changes, min_confidence_to_submit
    • Maintainer Tools: stale config, branch cleanup
    • Branch Enforcement, Merge Queue, AI Review Gate, Trust
  • Settings — Dry Run Mode toggle (yellow banner when active)
  • Change History — Audit trail of all config changes with:
    • Action badges (set, patch, delete, restore)
    • Before/after diff summary
    • Restore buttons to re-apply previous versions

Decisions (/decisions) ​

Worker decision log with full search/filter:

  • Free-text search — Search across decision reasons
  • Source filter — Filter by worker (ci_heal, triage, ai_review, etc.)
  • Decision filter — acted, skipped, dry_run, blocked
  • Pillar filter — 9 pillar dropdown
  • Date range — From/to date pickers
  • Expandable detail — Metadata, config_used JSON, conditions
  • Export — JSON and Markdown audit bundles
  • Pagination — Server-side paginated results

Dry-Run Proof (/dry-run) ​

Non-mutating evaluation evidence:

  • Safety-first labels — "Would have", "Skipped mutation", never implies execution
  • Decision pinned to dry_run — Cannot view executed actions here
  • Filters — Repo, pillar, source, target type, date range, free-text
  • Expandable detail — Planned reason, config used, metadata, conditions
  • Deep links — Links to decision log and managed actions
  • Export — JSON and Markdown audit bundles

Waivers (/waivers) ​

Policy waiver visibility across all repos:

  • Global view — All repos, not just one at a time
  • Status computation — Active, expiring (within 7 days), expired
  • Filters — Repo, pillar, scope, status, free-text search
  • Expandable detail — Metadata, reason, decision linkage
  • Decision linkage — Links to decision log filtered by matching pillar+repo+skipped
  • Grant/revoke — Create and revoke waivers (time-limited exceptions)
  • Export — JSON audit bundle

Activity (/activity) ​

Unified action feed across 9 source tables:

  • Source filter bar — Filter by action type
  • Color-coded badges — Per-source coloring
  • Pagination — Server-side paginated results

Policy Preview (/policy-preview) ​

Non-mutating policy analysis workflow:

  • YAML editor — Paste proposed .gitwire.yml for analysis
  • Validation panel — Errors, warnings, enabled pillars, dry-run state, risky settings with mitigation badges, collapsible normalized config
  • Simulation panel — Replay proposed policy against historical events; summary cards (Considered, Would act, Would skip, Dry-run, Block, Unsupported); per-event expandable rows with conditions, would_do, and original decision
  • Impact comparison panel — Compare current vs proposed policy; summary cards for dry-run change, pillar changes, risk changes; per-event impact table with current → proposed decision transitions
  • Recommendations panel — Deterministic guardrail recommendations grouped by severity (critical/warning/info); each card includes reason, suggested change, config path, and evidence chips
  • Safety banner — All operations are non-mutating: no config saves, no queue jobs, no GitHub writes
  1. Paste proposed .gitwire.yml and click Validate
  2. Review risks, warnings, and enabled pillars
  3. Select a repo and click Run simulation to see historical impact
  4. Select the same repo and click Compare impact to see what changes vs current policy
  5. Optionally select a repo and click Generate recommendations for rollout guardrails
  6. Roll out with dry-run first when enabling risky or newly permissive automation

Readiness (/readiness) ​

Rollouts (/rollouts) ​

Controlled policy lifecycle dashboard:

  • List view — All rollout plans with status badges, repo, dry-run indicator, risk counts, and actor trail
  • Filters — Status (8 states), repo full name
  • Detail view — Lifecycle timeline, evidence cards (validation, simulation, diff, recommendations), audit trail (all actors with timestamps and reasons)
  • Policy snapshots — Collapsible redacted JSON blocks (proposed, previous, replaced)
  • Rollback evidence — Config hashes and capture booleans
  • Actions panel — State-driven buttons showing only valid next actions
  • Confirmation modal — Actor input, reason textarea, consequence warnings
    • Promote/rollback: amber warning (writes policy)
    • Reject/cancel: red warning (terminal)
    • Approve: explains no policy written yet
  • Critical recommendation acknowledgement — Checkboxes for critical recs on approve

State-driven actions ​

StateActions
draftcancel
validatedcancel
review_readyapprove, reject, cancel
approvedpromote, cancel
promotedrollback
terminal(none)

Readiness (/readiness) — original ​

Fleet-wide repo readiness scores:

  • Fleet overview — All repos scored out of 100, sorted by score
  • Per-repo detail — 9 weighted checks with pass/fail status
  • Checks — .gitwire.yml present, webhooks active, branch protection, config synced, etc.

→ Configuration

Released under the MIT License.