You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Evolve the organization-level Intelligence portal from a directory/summary aggregator into a low-cognitive-load fleet control plane that composes the same normalized evidence used by repository-level Repository Intelligence.
The portal should let an organization architect answer "where are we?" without opening 29 repository dashboards, while always preserving drill-down to the canonical repository evidence.
This is a follow-up/expansion of #15, not a replacement for repository-level Intelligence.
Repositories — searchable fleet directory with lifecycle/class/profile.
The exact navigation should reuse the Repository Intelligence information architecture where practical rather than creating a second incompatible vocabulary.
Progressive disclosure
Optimize the default experience for quick architectural orientation:
compact organization overview;
exceptions/blockers/stale evidence;
repositories needing attention;
domain summaries;
drill-down into repository Intelligence;
direct source/evidence links when deeper inspection is needed.
Do not place every metric on one giant dashboard.
Aggregation semantics
For every organization-level object preserve:
source organization/repository;
represented revision;
snapshot/generated time;
freshness;
state/uncertainty;
source module/contract version;
deep link to repository-level view;
deep link to canonical evidence when authorized.
Unknown, unavailable, stale, partial, blocked, exempt, and not-applicable must remain distinct.
Audits
The organization audit view should support:
latest audit coverage by repository/profile;
unaudited/stale/partial repositories;
confirmed findings by severity/classification;
feature opportunities and experimental opportunities as separate categories;
recurring cross-repository patterns;
organization audit campaigns/runs;
candidate issue materialization status.
Do not turn speculative ideas into organization-health failures.
Hygiene
The organization Hygiene view should reuse Observatory #5 and allow:
rows/columns by repository and contract domain;
filters by repository class/profile/lifecycle;
explicit migration/advisory states;
deep links to repository /hygiene/;
prioritization by missing/stale/blocked evidence without pretending every contract applies everywhere.
Identity / decisions / other modules
Aggregate summary metadata only. Detailed Brand Kits, ADR prose, audit reports, roadmaps, etc. remain owned by the repository or their canonical semantic owner.
The portal should answer "which repository needs attention and why?" and then link out instead of copying every source document into one mega-store.
Data-source evolution
#15 initially permits fetching repository summary.json endpoints. Preserve that simple public path where useful, but allow the mature implementation to consume a deterministic organization snapshot produced by Observatory/Relay when that avoids N browser requests, CORS fragility, or inconsistent schema versions.
No browser token, authenticated GitHub scraping, database, or hosted backend is required for the static public implementation.
Privacy
Public portal output must not reveal private repository existence, counts, names, issues, paths, findings, or inferred gaps.
Private/local organization projections may compose authorized private snapshots separately.
Visibility classification is enforced before aggregation, not merely hidden with CSS.
Outcome
Evolve the organization-level Intelligence portal from a directory/summary aggregator into a low-cognitive-load fleet control plane that composes the same normalized evidence used by repository-level Repository Intelligence.
The portal should let an organization architect answer "where are we?" without opening 29 repository dashboards, while always preserving drill-down to the canonical repository evidence.
This is a follow-up/expansion of #15, not a replacement for repository-level Intelligence.
Product principle
The organization UI is a projection. It must never become a second source of truth.
Initial organization views
Provide modular organization-level views/tabs that aggregate safe repository summaries for domains such as:
egohygiene/observatory#5;egohygiene/observatory#20;The exact navigation should reuse the Repository Intelligence information architecture where practical rather than creating a second incompatible vocabulary.
Progressive disclosure
Optimize the default experience for quick architectural orientation:
Do not place every metric on one giant dashboard.
Aggregation semantics
For every organization-level object preserve:
Unknown, unavailable, stale, partial, blocked, exempt, and not-applicable must remain distinct.
Audits
The organization audit view should support:
Do not turn speculative ideas into organization-health failures.
Hygiene
The organization Hygiene view should reuse Observatory #5 and allow:
/hygiene/;Identity / decisions / other modules
Aggregate summary metadata only. Detailed Brand Kits, ADR prose, audit reports, roadmaps, etc. remain owned by the repository or their canonical semantic owner.
The portal should answer "which repository needs attention and why?" and then link out instead of copying every source document into one mega-store.
Data-source evolution
#15 initially permits fetching repository
summary.jsonendpoints. Preserve that simple public path where useful, but allow the mature implementation to consume a deterministic organization snapshot produced by Observatory/Relay when that avoids N browser requests, CORS fragility, or inconsistent schema versions.No browser token, authenticated GitHub scraping, database, or hosted backend is required for the static public implementation.
Privacy
Acceptance criteria
Dependencies / related
egohygiene/relay#27egohygiene/observatory#5egohygiene/observatory#20/audits/:egohygiene/relay#74/hygiene/:egohygiene/relay#75egohygiene/observatory#17Non-goals