What it is

The AI Resolution Hub is the fix queue for Agentic Foundation. It gathers every Gap, Partial, and Blocked criterion across all connected repos and sorts them so the highest-impact work floats to the top.

Use it when you do not want to open fifteen repo detail views to decide what to fix next.

How prioritization works

Items are ordered roughly like this:

1. 1
   Severity first

   High gaps appear above Medium and Low at the same repo.

   Example: “Add agent instructions” (High) on payments-api ranks above “Add EditorConfig” (Medium) on the same repo.

2. 2
   Dimension weight

   Gaps in Agentic Readiness and Test Coverage outrank equal-severity gaps in lighter pillars.

   Example: A High gap in Agentic Readiness on web-checkout can appear above a High gap in Token Efficiency on handbook.

3. 3
   Repository score

   Fixes on lower-scoring repos surface higher when severity and dimension are tied, because they move the portfolio needle more.

   Example: Two identical High gaps: one on payments-api at 54, one on handbook at 82. The API repo item tends to rank first.

Reading a queue item

Each row includes:

Field	What you see	Example
Repository	Source repo	payments-api
Dimension	Which of the nine pillars	Agentic Readiness
Severity	High / Medium / Low	High
Description	Plain-language finding	“No agent instruction file (AGENTS.md or equivalent)”
Remediation hint	Concrete next step	“Add AGENTS.md with copy-pasteable build, test, and lint commands”

Click through to the repo detail view if you need the full criterion list or score history.

Example: a Monday morning triage

Acme Engineering has three repos after the weekend scans:

Priority	Repo	Item	Why it’s near the top
1	payments-api	Add AGENTS.md	High severity + highest-weight dimension + low overall score
2	payments-api	CI does not run tests	High severity + Test Coverage weight
3	web-checkout	Add .cursorignore	High on Token Efficiency; repo already at 68
4	handbook	Expand README setup section	Medium severity on a repo that is already Excellent

The lead assigns items 1–2 to the API team, knocks out item 3 in a small PR, and leaves handbook for later.

Common fix patterns

These categories show up often. One PR can clear multiple queue rows.

Documentation & agent context

Fix	Files	Dimensions helped
Agent onboarding doc	AGENTS.md with npm test, npm run build, folder map	Agentic Readiness, Codebase Hygiene, Token Efficiency
Contributor guide	CONTRIBUTING.md, expanded README.md	Codebase Hygiene
Cursor setup	.cursor/rules/, .cursorignore	Agentic Readiness, Token Efficiency

Example: Merging AGENTS.md into payments-api removed three hub rows and moved Agentic Readiness from 41% to 68% on the next scan.

CI/CD

Fix	Files	Dimensions helped
Run tests in CI	.github/workflows/ci.yml with a test job	Test Coverage, Code Quality
Lint gate	npm run lint step that fails the workflow	Code Quality
Build artifact	Docker build stage in CI	Delivery Readiness

Example: Adding npm test to CI on web-checkout closed a High Test Coverage gap without touching application code.

Contracts & governance

Fix	Files	Dimensions helped
API spec	openapi.yaml or docs/openapi.yaml	API & Contract Design
Ownership	CODEOWNERS or .github/CODEOWNERS	Governance, Codebase Hygiene
Security policy	SECURITY.md, branch protection on main	Governance

Token efficiency

Fix	What to do	Dimensions helped
Ignore noise	.cursorignore excluding dist/, coverage/, node_modules/	Token Efficiency, Agentic Readiness
Split docs	Move long design notes from README into docs/architecture/	Token Efficiency
Tighten rules	Path-scoped .cursor/rules instead of always-on walls	Token Efficiency

Closing the loop

After you ship a fix:

1. 1
   Merge to the default branch

   Scans read the repo snapshot at scan time. Open PRs do not count until merged (unless you manually re-scan a feature branch, which is uncommon).

2. 2
   Re-scan

   From Agentic Foundation, click Re-scan on the repo row or use Analyze from the list.

3. 3
   Wait for completion

   Usually 5–15 minutes. The hub row disappears when the criterion moves to Met or N/A.

4. 4
   Check score history

   Open the repo detail view to confirm the overall and dimension scores moved the way you expected.

If an item stays on the hub after re-scan, read the remediation hint again: sometimes the file path, command wording, or CI job name must match what the rubric expects (e.g. tests must run in CI, not only locally).

When the hub is empty

Either every connected repo is in good shape (rare on first scan), or scans have not finished yet. Check for repos stuck in running or failed status on the main Agentic Foundation list.

Scheduled scans about every 24 hours will pick up merged fixes even if nobody clicks Re-scan, but manual re-scan is faster when you are actively working the queue.

Next

• Readiness scores
• The 9 dimensions
• Agentic Foundation overview