Static GitHub Pages delivery
Evidence: Briefing builder — scripts/build-briefing.mjs generates src/data/briefing.json from public-safe static sources.
Static AgentSite · product-cockpit recipe · no-payment mode
A static chief-of-staff brief for the next agent run
The page turns requests, roadmap items, run history, proof sites, and search metadata into a compact control brief for autonomous maintenance decisions.
The page explains the user-facing workflow first, then gives reviewers the operating context, repo-local proof, and approval boundaries needed to evaluate the next step safely.
Operating context
What decision this page helps visitors make
A static AgentSite chief-of-staff briefing for autonomous maintenance decisions.
Create a static GitHub Pages site that demonstrates a chief-of-staff briefing generated from public-safe AgentSite data: requests, roadmap, runs, atlas, and search index. It should show top three next actions, risk flags, stale or blocked work, recent verified wins, and the recommended recipe to apply next without live monitoring, private tickets, raw chats, auth, analytics, or a database.
Copy evidence strip
Every strong line gets paired with a verifiable artifact.
Public-safe chief-of-staff briefing
Evidence: Briefing QA — scripts/check-briefing.mjs validates top actions, risk flags, wins, recommendation, and source counts.
Repo-local QA scripts validate briefing records
Evidence: Briefing UI — src/components/ChiefOfStaffBriefing.astro renders the executive panel and source evidence links.
Payment disabled / no-payment mode
Evidence: Briefing builder — scripts/build-briefing.mjs generates src/data/briefing.json from public-safe static sources.
Artifact gallery
Browse the proof future agents can maintain.
Search and filter static repo artifacts, plans, QA evidence, deploy records, and recipes without adding a backend or live tracking.
Showing 9 artifacts
Contract verified
Site contract
Defines mission, required sections, allowed claims, approval boundaries, and live verification expectations.
contractsguardrailsagents
Inspect artifact Contract verified
Brand contract
Captures tone, visual posture, accessibility posture, and language future agents should avoid.
brandvoiceguardrails
Inspect artifact QA verified
QA command suite
Runs contract, claims, SEO, link, artifact, build, and visual checks before handoff.
qabuildverification
Inspect artifact Deploy verified
Deploy verifier
Checks the live GitHub Pages URL for stable expected text after deployment.
deploygithub-pagesverification
Inspect artifact Plan planned
Initial build plan
Records the original brief, assumptions, non-goals, acceptance criteria, and verification commands.
planhandoffagents
Inspect artifact Recipe verified
Recipe registry
Documents reusable archetypes and section recipes future agents can score before applying.
recipespatternsagents
Inspect artifact Evidence static
Briefing builder
scripts/build-briefing.mjs generates src/data/briefing.json from public-safe static sources.
proofevidence
Inspect artifact Evidence static
Briefing QA
scripts/check-briefing.mjs validates top actions, risk flags, wins, recommendation, and source counts.
proofevidence
Inspect artifact Evidence static
Briefing UI
src/components/ChiefOfStaffBriefing.astro renders the executive panel and source evidence links.
proofevidence
Inspect artifact Local-first roadmap
A static improvement board agents can keep current.
Filter the generated roadmap, then try local-only status changes. Browser overrides stay on your device and never imply repo changes or live workflow automation.
Showing 6 roadmap items
Now
highagent
Tighten recipe selection
Keep selector rules deterministic as more section recipes compose together.
Next: Add a regression config whenever selector behavior changes.
highagent
Briefing builder follow-up
scripts/build-briefing.mjs generates src/data/briefing.json from public-safe static sources.
Next: Keep this proof item linked to a real repo artifact before strengthening copy.
Next
highagent
Local-first user overrides
Let visitors locally move roadmap cards without implying shared task updates.
Next: Use localStorage for browser-only status overrides and keep durable changes in JSON.
mediumagent
Briefing QA follow-up
scripts/check-briefing.mjs validates top actions, risk flags, wins, recommendation, and source counts.
Next: Keep this proof item linked to a real repo artifact before strengthening copy.
Later
mediumagent
Static intelligence index
Generate precomputed indexes that make static sites feel more queryable.
Next: Prototype a JSON index from recipe, artifact, and roadmap metadata.
Blocked
mediumhuman
Autonomous maintenance cadence
Define when recurring agents may propose changes without adding notification noise.
Next: Approve a quiet schedule and escalation boundaries before automation.
Static intelligence pack
Search the site brain without a backend.
A precomputed JSON index spans recipes, artifacts, roadmap items, contracts, and plans. It makes the static page feel queryable while staying public-safe and GitHub Pages compatible.
Showing 28 indexed entries
Site contract
Public source of truth for site mission, required sections, allowed claims, and approval boundaries.
Open indexed targetREADME handoff
Repository overview, selected recipes, live URL, QA commands, and agent maintenance notes.
Open indexed targetAgent Run Ledger
A static orchestration ledger for agent-maintained sites: decisions, delegated
Open indexed targetAgentSite Atlas
A static directory of generated AgentSite proof sites, including live links,
Open indexed targetArtifact Gallery
A static, searchable, filterable gallery for repo artifacts, QA evidence,
Open indexed targetChief-of-staff briefing
Generates a public-safe executive briefing panel from static AgentSite data:
Open indexed targetCopy Evidence Strip
A composable AgentSite pattern that pairs the strongest visible claims with
Open indexed targetEditorial Ledger
A full-page AgentSite archetype for credibility-first narrative pages that read
Open indexed targetFeature Request Inbox
A static intake queue for future AgentSite improvements with status, priority,
Open indexed targetProduct Cockpit
A reusable AgentSite pattern for turning a product, service, or internal initiative
Open indexed targetRoadmap Board
A static local-first improvement board for AgentSite repos. It renders a
Open indexed targetSearch Index
A static intelligence pack that precomputes a public-safe JSON index over
Open indexed targetSite contract
Defines mission, required sections, allowed claims, approval boundaries, and live verification expectations.
Open indexed targetBrand contract
Captures tone, visual posture, accessibility posture, and language future agents should avoid.
Open indexed targetQA command suite
Runs contract, claims, SEO, link, artifact, build, and visual checks before handoff.
Open indexed targetDeploy verifier
Checks the live GitHub Pages URL for stable expected text after deployment.
Open indexed targetInitial build plan
Records the original brief, assumptions, non-goals, acceptance criteria, and verification commands.
Open indexed targetRecipe registry
Documents reusable archetypes and section recipes future agents can score before applying.
Open indexed targetBriefing builder
scripts/build-briefing.mjs generates src/data/briefing.json from public-safe static sources.
Open indexed targetBriefing QA
scripts/check-briefing.mjs validates top actions, risk flags, wins, recommendation, and source counts.
Open indexed targetBriefing UI
src/components/ChiefOfStaffBriefing.astro renders the executive panel and source evidence links.
Open indexed targetTighten recipe selection
Keep selector rules deterministic as more section recipes compose together. Add a regression config whenever selector behavior changes.
Open indexed targetLocal-first user overrides
Let visitors locally move roadmap cards without implying shared task updates. Use localStorage for browser-only status overrides and keep durable changes in JSON.
Open indexed targetStatic intelligence index
Generate precomputed indexes that make static sites feel more queryable. Prototype a JSON index from recipe, artifact, and roadmap metadata.
Open indexed targetAutonomous maintenance cadence
Define when recurring agents may propose changes without adding notification noise. Approve a quiet schedule and escalation boundaries before automation.
Open indexed targetBriefing builder follow-up
scripts/build-briefing.mjs generates src/data/briefing.json from public-safe static sources. Keep this proof item linked to a real repo artifact before strengthening copy.
Open indexed targetBriefing QA follow-up
scripts/check-briefing.mjs validates top actions, risk flags, wins, recommendation, and source counts. Keep this proof item linked to a real repo artifact before strengthening copy.
Open indexed targetInitial site build plan: Chief of Staff Briefing
Initial site build plan: Chief of Staff Briefing Request Create a static GitHub Pages site that demonstrates a chief of staff briefing generated from public safe AgentSite data: requests, roadmap, runs, atlas, and search
Open indexed targetChief-of-staff briefing
A generated control brief for the next agent decision.
Static chief-of-staff briefing generated from public AgentSite data: requests, roadmap, runs, atlas, and search index. Generated 2026-05-16. This is static public data, not live monitoring.
Top 3Next actions
Do these before starting broad new work.
- Move primary run ledger closer to top
Run-ledger proof is useful but buried below composed modules. Next: Prototype a run-ledger-first layout in a disposable generated repo.
request:req-001 - Generate a site-specific chief-of-staff briefing
Generate a compact static chief-of-staff briefing from atlas, runs, requests, roadmap, and search index data. Next: Publish proof and back-port generator support.
request:req-005 - Add collapsible module sections
Composed demo pages become too long when multiple section recipes are enabled. Next: Design a static details/summary pattern with visual QA coverage.
request:req-002
RiskFlags
What can slow the next autonomous run.
- Autonomous maintenance cadence
Roadmap item is blocked. Approve a quiet schedule and escalation boundaries before automation.
- Layer 8: chief-of-staff briefing
Run status is draft; deploy or verification may still need curation.
VerifiedWins
Recent proof the system can keep compounding.
- Layer 7: feature request inbox
Added public-safe feature-request inbox with status, priority, acceptance criteria, recipe candidates, verification plan, and next action.
- Chief of Staff Briefing launch run
A static AgentSite chief-of-staff briefing for autonomous maintenance decisions.
- Layer 5: AgentSite Atlas
Added searchable public directory of generated proof sites and recipe coverage.
Recommendationagent-run-ledger
Recommended recipe to apply next.
Most relevant candidate in the current public request/roadmap data: agent-run-ledger.
Evidence sources used for this static brief
Feature request inbox
A static intake queue for the next agent run.
Requests are public-safe JSON records with status, priority, acceptance criteria, recipe candidates, verification plan, and next action.
Showing 5 requests
highready
Move primary run ledger closer to top
Run-ledger proof is useful but buried below composed modules.
- Source
- visual review
- Next
- Prototype a run-ledger-first layout in a disposable generated repo.
Acceptance
- #runs appears before secondary proof modules
- top CTA jumps to runs
- visual QA passes
Verification
- npm run qa:full
- browser visual review
mediumtriage
Add collapsible module sections
Composed demo pages become too long when multiple section recipes are enabled.
- Source
- atlas/run-ledger visual review
- Next
- Design a static details/summary pattern with visual QA coverage.
Acceptance
- secondary modules can collapse without hiding core content
- keyboard behavior works
- mobile remains readable
Verification
- npm run qa:full
- manual keyboard check
highdone
Chief of Staff Briefing launch request
A static AgentSite chief-of-staff briefing for autonomous maintenance decisions.
- Source
- Layer 7
- Next
- Use the inbox as input for briefing generation.
Acceptance
- requests JSON validates
- request cards filter by status and priority
- QA exercises request search/filter
Verification
- npm run check:requests
- npm run qa:full
- npm run verify:deploy
mediumbacklog
Track failed experiments explicitly
Future agents need a clean place to record dead ends and why they were rejected.
- Source
- orchestration model
- Next
- Extend schemas after inbox proof is stable.
Acceptance
- failure states are visible
- reason and rollback note are required
- blocked language checks pass
Verification
- npm run check:runs
- npm run check:requests
highin_progress
Generate a site-specific chief-of-staff briefing
Generate a compact static chief-of-staff briefing from atlas, runs, requests, roadmap, and search index data.
- Source
- Layer 8
- Next
- Publish proof and back-port generator support.
Acceptance
- briefing is generated from public data
- top three actions are explicit
- no live monitoring claims
Verification
- npm run build:briefing
- npm run check:briefing
- npm run qa:full
- npm run verify:deploy
Agent run ledger
A static orchestration trail future agents can inspect.
Each run records the layer, decision, delegated work, verification checks, deploy evidence, and the next action. It is a public handoff ledger, not live telemetry.
Showing 8 runs
Layer 1verified
Layer 1: archetype divergence
Added editorial-ledger and visual divergence guard.
- Date
- 2026-05-15
- Orchestrator
- Hermes Agent
- Commit
- eca9b19c39cdbc35a20ffd52e097232998ac6fcc
Delegated work
- Generator patch
- Editorial proof publish
- Visual review
Checks
- npm run qa:full
- npm run check:visual-divergence
- npm run verify:deploy
Next: Add richer static interactions.
Deploy evidenceLayer 2verified
Layer 2: artifact gallery
Added searchable/filterable artifact gallery with URL tag state.
- Date
- 2026-05-16
- Orchestrator
- Hermes Agent
- Commit
- 9ce018487f7d4706db063ce14d4b5aca7b3675c7
Delegated work
- Recipe implementation
- Proof site publish
- Final review
Checks
- npm run qa:full
- npm audit --audit-level=moderate
- npm run verify:deploy
Next: Add local-first roadmap state.
Deploy evidenceLayer 3verified
Layer 3: roadmap board
Added JSON-backed roadmap board with localStorage status overrides.
- Date
- 2026-05-16
- Orchestrator
- Hermes Agent
- Commit
- a266bd924d509e4ce6365628baf0ee0a352a9c13
Delegated work
- Local-first component
- Roadmap QA
- Proof site publish
Checks
- npm run qa:full
- npm run check:roadmap
- npm run verify:deploy
Next: Add precomputed static intelligence index.
Deploy evidenceLayer 4verified
Layer 4: search index
Added public-safe static search over contracts, recipes, artifacts, roadmap, and plans.
- Date
- 2026-05-16
- Orchestrator
- Hermes Agent
- Commit
- 2b632ce9dd64e454291bb06ecd637d2adbf05b25
Delegated work
- Index builder
- Search UI
- Visual QA
Checks
- npm run build:search-index
- npm run check:search-index
- npm run qa:full
Next: Add generated proof-site atlas.
Deploy evidenceLayer 5verified
Layer 5: AgentSite Atlas
Added searchable public directory of generated proof sites and recipe coverage.
- Date
- 2026-05-16
- Orchestrator
- Hermes Agent
- Commit
- 75a23145fd1d834560165feb7f5e9b6ff50f9878
Delegated work
- Atlas data
- Atlas UI
- Deploy verification
Checks
- npm run check:atlas
- npm run qa:full
- npm run verify:deploy
Next: Model the orchestrator/subagent maintenance loop as a static run ledger.
Deploy evidenceLayer 6verified
Chief of Staff Briefing launch run
A static AgentSite chief-of-staff briefing for autonomous maintenance decisions.
- Date
- 2026-05-16
- Orchestrator
- Hermes Agent
- Commit
- 43f0cfcfb9ea8ff45018bf5bae693e7665c2f3fd
Delegated work
- Run schema
- Ledger UI
- Run QA
Checks
- npm run check:runs
- npm run qa:full
Next: Add static request intake queue.
Deploy evidenceLayer 7verified
Layer 7: feature request inbox
Added public-safe feature-request inbox with status, priority, acceptance criteria, recipe candidates, verification plan, and next action.
- Date
- 2026-05-16
- Orchestrator
- Hermes Agent
- Commit
- 5fff83b9116db021be83480e4a1ff7e4ea7afbcd
Delegated work
- Request schema
- Inbox UI
- Request QA
Checks
- npm run check:requests
- npm run qa:full
- npm run verify:deploy
Next: Generate a chief-of-staff briefing from requests, roadmap, runs, atlas, and search index.
Deploy evidenceLayer 8draft
Layer 8: chief-of-staff briefing
Planned generated static chief-of-staff briefing for top actions, risk flags, stale work, wins, and next recipe recommendation.
- Date
- 2026-05-16
- Orchestrator
- Hermes Agent
- Commit
- pending
Delegated work
- Briefing builder
- Executive panel UI
- Briefing QA
Checks
- npm run build:briefing
- npm run check:briefing
- npm run qa:full
Next: Publish proof site and verify deploy.
AgentSite Atlas
A generated directory of proof sites.
The atlas turns prior generated sites into a navigable system: live links, repo links, recipe coverage, deploy evidence, and the next useful improvement.
Showing 7 proof sites
Layer 0verified
Chief of Staff Briefing
A static AgentSite chief-of-staff briefing for autonomous maintenance decisions.
- Archetype
- default-landing
- Repo
- castiliad/agentsite-chief-briefing-generated-20260516022220
Next: Continue adding verified static-site capability layers.
Layer 1verified
Product Cockpit Proof
Dark cockpit archetype proof for operator/product-reviewer audiences.
- Archetype
- product-cockpit
- Repo
- castiliad/agentsite-cockpit-generated-20260515212318
Next: Use for technical product/control-plane sites.
Layer 1verified
Editorial Ledger Proof
Light editorial memo/ledger archetype proving visual divergence from cockpit layouts.
- Archetype
- editorial-ledger
- Repo
- castiliad/agentsite-editorial-ledger-generated-20260515234524
Next: Use for credibility-first narratives and provenance-heavy pages.
Layer 2verified
Evidence Strip Proof
Claim-to-artifact section recipe for grounding marketing copy in verifiable evidence.
- Archetype
- default-landing
- Repo
- castiliad/agentsite-evidence-strip-generated-20260515231333
Next: Keep proof copy tied to real repo artifacts.
Layer 2verified
Artifact Gallery Proof
Searchable/filterable proof browser backed by static artifact JSON.
- Archetype
- product-cockpit
- Repo
- castiliad/agentsite-artifact-gallery-generated-20260516001714
Next: Use for proof-heavy handoff sites.
Layer 3verified
Roadmap Board Proof
Local-first roadmap mini app with JSON data, filters, columns, and localStorage status overrides.
- Archetype
- product-cockpit
- Repo
- castiliad/agentsite-roadmap-board-generated-20260516005337
Next: Use for public next-step queues and safe agent-maintained roadmaps.
Layer 4verified
Search Index Proof
Static intelligence pack with precomputed search over public contracts, recipes, artifacts, roadmap items, and plans.
- Archetype
- product-cockpit
- Repo
- castiliad/agentsite-search-index-generated-20260516011158
Next: Use when a site needs queryable public repo intelligence without backend search.
Audience named in the brief
Built for the people who need the workflow explained clearly.
- AgentSite chief-of-staff agents
- Autonomous orchestrators
- Technical reviewers evaluating static maintenance loops
Cockpit panels
Status, workflow, proof, and constraints in one scan.
The chief-of-staff view starts with decisions
The generated briefing surfaces the few actions and risks an orchestrator should review before delegating work.
Static signals replace vague status chatter
Requests, roadmap, runs, atlas, and search metadata are transformed into a curated public-safe operating summary.
No live monitoring or private intake
The panel is generated from committed static data, not private tickets, raw chats, analytics, telemetry, auth, or a database.
Proof artifacts
Evidence a future maintainer can verify.
Briefing builder
scripts/build-briefing.mjs generates src/data/briefing.json from public-safe static sources.
Briefing QA
scripts/check-briefing.mjs validates top actions, risk flags, wins, recommendation, and source counts.
Briefing UI
src/components/ChiefOfStaffBriefing.astro renders the executive panel and source evidence links.
Boundaries
What is not claimed here.
Not claimed: fake customer endorsements.
Not claimed: fake metrics.
Not claimed: unsupported guarantees.
Approval required: Adding analytics, cookies, tracking pixels, or third-party forms.
Approval required: Adding payment links or enabling payment mode.
Approval required: Indexing private tickets, raw chats, credentials, secrets, or unpublished user data.
Safe next step
Review briefing
Use the cockpit panels, repository contracts, and QA commands before changing claims, scope, payments, analytics, or deployment settings.