Agency OS · Architecture & Training
Sections

Indexed repo docs

Source Library.

Raw ADRs, runbooks, decisions, and repo-owned skill files indexed for technical users.

Documents

  • Admin Registry Source Asset Write Semantics decisions/admin-registry-source-asset-write-semantics.md Admin Registry Source Asset Write Semantics Date: 2026-06-05 Status Accepted and implemented. Context The Admin Registry Business Unit Source Assets endpoint accepted both PUT and PATCH , but both methods ran the same partial upsert loop. Omitted Source Assets
  • Aggregate Performance Reporting and Clarification Loop Guard decisions/aggregate-performance-reporting-and-clarification-loop-guard.md Aggregate Performance Reporting and Clarification Loop Guard Date: 2026-05-20 Decision Multi-channel report requests such as "aggregated April report for Plunkett including SEO, Google Ads and Meta" must not route to the first channel keyword found. They now r
  • Cloudflare Agents SDK v0.14 And Workflow Schedules decisions/cloudflare-agents-sdk-v014-and-workflow-schedules.md Cloudflare Agents SDK v0.14 And Workflow Schedules Date: 2026-06-04 Status: Accepted with staged adoption Task: Review Remediation: Evaluate Cloudflare Agents SDK v0.14 and scheduled Workflows adoption https://app.asana.com/1/1079477897941/project/121479888426
  • Context Refresh Resumable Box Scan decisions/context-refresh-resumable-box-scan.md Context Refresh Resumable Box Scan Date: 2026-07-02 Status Accepted and implemented. Context The scheduled context refresh scans the configured Box folder tree before it diffs and ingests candidate context files. Plunkett Homes now points at the JWH GROUP root
  • Expanded Forensic Plunkett SEO Report Decisions decisions/forensic-plunkett-seo-report.md Expanded Forensic Plunkett SEO Report Decisions Goal Complete the expanded internal forensic SEO report for Plunkett hosted Team Agent execution. Current Baseline - Asana tasks assigned to Jessie in Jessie Intake can trigger hosted Team Agent execution. - Plun
  • GA4 Reporting Skill Decisions decisions/ga4-reporting-skill.md GA4 Reporting Skill Decisions Date: 2026-05-21 Decision Add ga4-reporting as a hosted Team Agent skill for internal analytics reporting. The first implementation slice supports an Asana task such as: The report should be highly digestible for consultants. Majo
  • General Jessie Tasks and Follow-Up Comments decisions/general-jessie-task-and-follow-up-comments.md General Jessie Tasks and Follow-Up Comments Date: 2026-05-20 Decision Jessie can accept assigned Asana tasks that do not map to a specific reporting or audit skill by falling back to general-task . The first hosted implementation is intentionally bounded: it r
  • General Jessie Workbench decisions/general-jessie-workbench.md General Jessie Workbench Date: 2026-05-21 Decision Jessie should not require a bespoke hosted skill or hardcoded lane for every consultant request. The target product shape is a general workbench: 1. Read the Asana task brief and recent consultant comments. 2.
  • Google Ads Reporting Channel Skill decisions/google-ads-reporting-skill.md Google Ads Reporting Channel Skill Status Started 2026-05-20. Decision Create a reusable Agency OS channel skill for Google Ads reporting, separate from the SEO reporting and technical SEO audit skills. The first implementation slice is reporting-only: - Defin
  • Harvest to Dashboard Client Backfill Reconciliation decisions/harvest-active-client-dashboard-backfill.md Harvest to Dashboard Client Backfill Reconciliation Date: 2026-07-08 Status Accepted; implemented build tasks 1-6, tests green . First-live prod rollout pending explicit operator go. Context Only a subset of the agency's active clients were configured in the a
  • Hosted Reporting Live Evidence Guard decisions/hosted-reporting-live-evidence-guard.md Hosted Reporting Live Evidence Guard Date: 2026-06-04 Status Accepted and implemented. Context Jessie's hosted reporting path still had deterministic Plunkett fixture builders available from the workflow package. Those fixtures are useful for package contracts
  • Local Agent Worktree Workflow decisions/local-agent-worktree-workflow.md Local Agent Worktree Workflow Date: 2026-05-24 Decision Local Agency OS development agents should use short-lived, task-scoped Git worktrees for non-trivial implementation work. The root checkout at C:\Users\StuartWatson\Box\Living Online\agency-os is the coor
  • MCP Tool Plane Hand-Rolled Server decisions/mcp-tool-plane-hand-rolled-server.md MCP Tool Plane Hand-Rolled Server Date: 2026-06-10 Decision Implement the apps/mcp-tools MCP Streamable HTTP server as a hand-rolled JSON-RPC 2.0 handler with no new npm dependencies. Context ADR-0019 introduces a scoped-token tool plane so Claude Agent SDK su
  • Meta Ads Reporting Skill Decisions decisions/meta-ads-reporting-skill.md Meta Ads Reporting Skill Decisions Date: 2026-05-20 Decision Add meta-ads-reporting as a hosted Team Agent skill that mirrors the Google Ads reporting flow: - Asana task request such as Please do a Meta Ads report for Plunkett Homes last month . - Live Meta Ma
  • skill r2 sync and hydration contract decisions/skill-r2-sync-and-hydration-contract.md Skill R2 Sync and Hydration Contract Status: Accepted Date: 2026-07-07 Context Execution-plane skills are stored in Git under skills/ , but runs hydrate skill documents from the CLIENT CONTEXT BUCKET R2 bucket under skills/ / . A missing R2 object for skills/t
  • Strategy Planning Skill And Harvest Budget Capacity decisions/strategy-planning-skill-and-harvest-budget.md Strategy Planning Skill And Harvest Budget Capacity Date: 2026-05-24 Decision Jessie's 90-day strategy planning capability is a first-class skill, not a one-off runner path. The skill contract lives in skills/strategy-planning/SKILL.md and the runtime policy d
  • Subagent Delegation for Agent Work decisions/subagent-delegation-for-agent-work.md Subagent Delegation for Agent Work Date: 2026-05-24 Decision Development agents should explicitly consider subagent delegation at the start of every non-trivial task, and should use subagents when they can split work into useful independent tracks. Good delega
  • Technical SEO Analyzer Tool Plane decisions/technical-seo-analyzer-tool-plane.md Technical SEO Analyzer Tool Plane Status Accepted on 2026-07-06. Context The technical SEO execution-plane port added a container-local analyzer CLI under apps/execution-plane/analyzer/ and copied shared workflow packages into the container image. The executio
  • Technical SEO Audit Proactive Crawl Parking decisions/technical-seo-audit-proactive-crawl-parking.md Technical SEO Audit Proactive Crawl Parking Date: 2026-07-03 Status Accepted and implemented. Context Long technical SEO Browser Run crawls can outlive a single platform-build control-plane Worker invocation. The existing graceful parking path already persists
  • Technical SEO Audit Synthesis And Context Promotion decisions/technical-seo-audit-synthesis-and-context-promotion.md Technical SEO Audit Synthesis And Context Promotion Date: 2026-06-04 Status: Accepted Context The SavvyWise technical SEO audit exposed two runtime gaps: - Cloudflare AI Gateway returned generated text in a result.output shape that the adapter did not unwrap,
  • Clean Agency OS Repository docs/adr/0001-clean-agency-os-repository.md Clean Agency OS Repository The Agency Operating System will be implemented in a new clean repository rather than by incrementally extending the current AIOps repository. The current repo is useful research for ingestion, reporting, JWH context, and early agent
  • ADR 0002: Hosted Platform Build Loop Control Plane docs/adr/0002-hosted-platform-build-loop-control-plane.md ADR 0002: Hosted Platform Build Loop Control Plane Date: 2026-05-15 Status: Accepted, corrected 2026-05-18 Context The first Platform Build Loop MVP is local and operator-driven. It can validate the Asana board, assess task readiness, post Delegation Confirmat
  • ADR 0003: Hosted Team Agent Work Intake docs/adr/0003-hosted-team-agent-work-intake.md ADR 0003: Hosted Team Agent Work Intake Date: 2026-05-18 Status: Accepted Current operating note: ADR 0016 now makes prod the only live client data environment. References below to the "live dev" intake describe the historical first deployment posture, not the
  • ADR 0004: Hosted Team Agent Execution MVP docs/adr/0004-hosted-team-agent-execution.md ADR 0004: Hosted Team Agent Execution MVP Date: 2026-05-18 Status: Accepted Context Hosted Team Agent intake can now notice Jessie Intake tasks, verify Jessie assignment, create a D1 Team Agent Run, assess readiness, and post a Delegation Confirmation. That pr
  • ADR 0005: Live Plunkett SEO Data Snapshot MVP docs/adr/0005-live-plunkett-seo-data-snapshot.md ADR 0005: Live Plunkett SEO Data Snapshot MVP Date: 2026-05-19 Status: Accepted Context Hosted Team Agent execution could produce a review-ready Plunkett SEO artifact from a synthetic fixture. To make the review materially useful, the runner needs to fetch rea
  • ADR 0006: Cloudflare Client Data Plane docs/adr/0006-cloudflare-client-data-plane.md ADR 0006: Cloudflare Client Data Plane Date: 2026-05-19 Status: Accepted Superseding note: ADR 0016 narrows live operation to the prod environment only until a dedicated development environment is explicitly designed. The fixture/dev environment model describe
  • ADR 0007: Jessie Teams Conversation Slice docs/adr/0007-jessie-teams-conversation-slice.md ADR 0007: Jessie Teams Conversation Slice Date: 2026-05-20 Status: Accepted Context Agency OS needs a way for staff to speak with Jessie in Microsoft Teams without turning Teams into a second workflow system of record. Existing decisions keep Asana as the syst
  • ADR 0007: Tracked Keyword Performance As Optional Primary SEO Rank Evidence docs/adr/0007-tracked-keyword-performance.md ADR 0007: Tracked Keyword Performance As Optional Primary SEO Rank Evidence Status: Accepted Search Console average position describes the broad query universe and can be distorted by long-tail queries, so it should not be the primary rank-performance KPI when
  • ADR 0008: Box Client Context Ingestion docs/adr/0008-box-client-context-ingestion.md ADR 0008: Box Client Context Ingestion Date: 2026-05-21 Status: Accepted Context Agency OS needs Jessie and future agents to use client documents, brand assets, meeting records, strategy notes, reports, and other context without giving agents unrestricted live
  • ADR 0009: Team Agent Runtime Extraction docs/adr/0009-team-agent-runtime-extraction.md ADR 0009: Team Agent Runtime Extraction Date: 2026-05-22 Status: Accepted Context The Cloudflare control-plane Worker has become the always-on runtime for Agency OS. It owns Asana webhook intake, queue consumers, Team Agent state transitions, execution leases,
  • ADR 0010: Team Agent Runtime Observability docs/adr/0010-team-agent-runtime-observability.md ADR 0010: Team Agent Runtime Observability Date: 2026-05-22 Status: Accepted Context Live Jessie task failures showed that the hosted runner had too little operational visibility. D1 run status could say running , queued , or failed , but operators could not q
  • ADR 0011: Live AI Gateway Model-Call Adapter docs/adr/0011-live-ai-gateway-model-call-adapter.md ADR 0011: Live AI Gateway Model-Call Adapter Date: 2026-05-22 Status: Accepted Context Jessie already carried AI Gateway configuration and model-call metadata, but hosted execution did not have a central live model-call adapter. ADR 0002 set the platform direc
  • ADR 0012: Major-Task Architecture Logging Hook docs/adr/0012-major-task-architecture-logging.md ADR 0012: Major-Task Architecture Logging Hook Date: 2026-05-24 Status: Accepted Context Agency OS already has a repository instruction that checks the Cloudflare Developer Changelog before platform, architecture, or implementation decisions. That hook is usef
  • ADR 0013: Asana-Sourced Jessie Scheduling and Standing Work docs/adr/0013-asana-sourced-jessie-scheduling-and-monitoring.md ADR 0013: Asana-Sourced Jessie Scheduling and Standing Work Date: 2026-05-24 Status: Accepted Implementation note 2026-05-24 : The foundation slice is implemented. Hosted Jessie intake reads Asana start/due/modified fields, parks future-start ready work as wai
  • ADR 0014: LLM-First Client-Context-Aware Team Agent Runtime docs/adr/0014-llm-first-client-context-aware-team-agent-runtime.md ADR 0014: LLM-First Client-Context-Aware Team Agent Runtime Date: 2026-05-24 Status: Accepted Context Team Agent reporting and insight skills were at risk of producing deterministic data summaries even when their skill contracts described LLM prompts. That mad
  • ADR 0015: R2-Backed Client Context Lake docs/adr/0015-r2-backed-client-context-lake.md ADR 0015: R2-Backed Client Context Lake Date: 2026-05-24 Status: Accepted Operating note: ADR 0016 now requires live Context Lake rows and R2 keys to use the prod logical environment. Any older fixture or dev context rows are historical only. Context The Box c
  • ADR 0016: Single Live Production Client Environment docs/adr/0016-single-live-production-client-environment.md ADR 0016: Single Live Production Client Environment Status: accepted Date: 2026-05-24 Context Agency OS had accumulated fixture , dev , and prod labels across the Client Data Plane, Admin Registry, Team Agent context resolution, scripts, and remote D1 rows. Se
  • ADR 0017: Approved Strategy Task Creation docs/adr/0017-approved-strategy-task-creation.md ADR 0017: Approved Strategy Task Creation Date: 2026-05-25 Status Accepted Context Jessie's 90-day strategy planning skill can produce a review-required strategy artifact with proposed delivery work, budget capacity, timing, and approval caveats. Consultants n
  • ADR 0018: Agency Brain Lean Context Runtime docs/adr/0018-agency-brain-lean-context-runtime.md ADR 0018: Agency Brain Lean Context Runtime Date: 2026-06-09 Status: Accepted Context Living Online consultants should be able to delegate work to Jessie through Asana without managing prompt packs, memory files, cache state, embeddings, or retrieval settings.
  • ADR-0019: Agent-Loop Execution Core for Team Agent Work docs/adr/0019-agent-loop-execution-core.md ADR-0019: Agent-Loop Execution Core for Team Agent Work Status: Accepted 2026-06-10 Date: 2026-06-10 Deciders: Stuart Watson Context Agency OS today executes Team Agent Jessie work through bespoke per-skill TypeScript pipelines inside the control-plane Worker:
  • ADR 0020: Source Asset Backed Google Ads MCP Evidence docs/adr/0020-google-ads-source-asset-backed-mcp-evidence.md ADR 0020: Source Asset Backed Google Ads MCP Evidence Status: accepted Date: 2026-06-19 Context The Team Agent execution path for Google Ads reporting had two conflicting ownership models: - Admin Registry source assets is the live system of record for client
  • ADR 0021: Autonomous Team Agent Client Onboarding docs/adr/0021-autonomous-team-agent-client-onboarding.md ADR 0021: Autonomous Team Agent Client Onboarding Status: accepted Date: 2026-06-25 Context Team Agent intake can receive valid Jessie work before a client has been configured in the live prod Admin Registry. Before this decision, the safe path was to persist
  • 0022. Staff admin dashboard v1 (STAFF ONLY) docs/adr/0022-staff-admin-dashboard.md 0022. Staff admin dashboard v1 STAFF ONLY Date: 2026-07-03 Status: accepted Context The dashboard system had drifted: a staff dashboard hardcoded to one fixture client, a freshness badge implying automation that did not exist, dead client-facing projection cod
  • 0023. Box webhook context ingest docs/adr/0023-box-webhook-context-ingest.md 0023. Box webhook context ingest Date: 2026-07-06 Status: accepted Context Consultants own each client's Box Agent Context folder. Before this change, Agency OS relied on scheduled context refresh as the main ingestion loop, so new context could wait for the n
  • ADR 0024: Durable Execution-Plane Result Landing and Watchdog Harvest docs/adr/0024-durable-execution-plane-result-harvest.md ADR 0024: Durable Execution-Plane Result Landing and Watchdog Harvest Date: 2026-07-07 Status: Accepted Context Execution-plane Team Agent runs were completed through a single awaited Worker call chain from the control-plane queue consumer into the execution-p
  • ADR 0025: Execution Plane Progress Observability docs/adr/0025-execution-plane-progress-observability.md ADR 0025: Execution Plane Progress Observability Status: accepted Date: 2026-07-07 Context Execution-plane technical-seo-audit runs could spend material wall-clock time without consultant-visible progress. The control plane recorded the service-binding call st
  • ADR-0026: Channel Analysis And Recommendation Pipeline docs/adr/0026-channel-analysis-recommendation-pipeline.md ADR-0026: Channel Analysis And Recommendation Pipeline Date: 2026-07-10 Status: Accepted for incremental implementation; external writes remain disabled. Context Jessie already has channel-specific reporting paths and an Asana-based consultant review loop, but
  • Agency OS Client Request Intake Note docs/agency-os-client-request-intake-note.md Agency OS Client Request Intake Note Client requests may enter Agency OS through a near-future HubSpot integration, but this does not create direct client access to Jessie or Asana. HubSpot Client Request Intake Slice A near-future operating slice that turns c
  • Agency OS Macro Architecture Proposal docs/agency-os-macro-architecture-proposal.html Agency OS Macro Architecture Proposal Living Online Agency OS Macro architecture proposal A plain-language proposal for the first useful Agency OS: staff allocate work in Asana, agents discuss the task, agents use client context and current data, and humans ap
  • Agency OS Operating Loops Note docs/agency-os-operating-loops-note.md Agency OS Operating Loops Note Agency OS should treat client performance improvement and agency operating leverage as two related but separate feedback loops. Client Performance Operating Loop Optimizes client outcomes using client goals, strategy, delivery ta
  • Agency OS Session Context - 18 May 2026 docs/agency-os-session-context-2026-05-18.md Agency OS Session Context - 18 May 2026 This note captures the working context from the architecture discussion so a restarted Codex session can pick up without re-litigating the same decisions. Session Intent Stuart is building Agency OS for Living Online. Th
  • Architectural Change Log docs/architecture/architectural-change-log.md Architectural Change Log This log is the short task-level index of material Agency OS architecture changes. It does not replace ADRs, decision files, runbooks, or CONTEXT.md ; it points staff to the durable records that changed during major tasks. Logging Rule
  • Jessie Asana-Sourced Scheduling and Standing Work Scope docs/architecture/jessie-asana-scheduled-monitoring-scope.md Jessie Asana-Sourced Scheduling and Standing Work Scope Date: 2026-05-24 Implementation foundation: 2026-05-24 Purpose This scope defines how Jessie should support future-scheduled work and recurring standing work while keeping Asana as the visible, consultant
  • Platform Build Loop MVP docs/architecture/platform-build-loop-mvp.md Platform Build Loop MVP Purpose The Platform Build Loop is the governed operating loop for changing the Agency Operating System itself. It turns Asana Platform Build Tasks into bounded builder-agent work, review artifacts, and human merge decisions. This MVP p
  • Agency OS Reconciliation Audit — 2026-06-15 docs/audit/2026-06-15-reconciliation.md Agency OS Reconciliation Audit — 2026-06-15 Read-only audit reconciling the pre-existing system against the new agent-loop ADR-0019 build. Six parallel investigations LLM paths, credentials, integrations, legacy→new gaps, runtime surface, branches+dashboards ,
  • Channel Analysis Pipeline Baseline docs/audit/2026-07-10-channel-analysis-baseline.md Channel Analysis Pipeline Baseline Date: 2026-07-10 Source State - Base: origin/main at 73b114268cd649f5f6f6aa0ec6c0fd5c14671657 . - Implementation branch: codex/channel-analysis-contracts . - Initial branch commit: ae5b2a03326fc736d24580ef89d5f44eb63e1178 . -
  • Worktree Reconciliation Audit docs/audit/2026-07-10-worktree-reconciliation.md Worktree Reconciliation Audit Date: 2026-07-10 Reference: origin/main at 73b114268cd649f5f6f6aa0ec6c0fd5c14671657 PR 102 Purpose Reconcile the C-drive Agency OS coordinator, managed worktrees, stale claims, and residual directories before expanding the Jessie
  • Claude_01 — Extract the Worker queue handler docs/Claude/Claude_01_extract_worker_queue_handler.md Claude 01 — Extract the Worker queue handler Priority: 1 highest operational risk Status: Proposed Owner: TBD Last updated: 2026-05-22 Why this matters apps/platform-build-control-plane/src/index.ts ../../apps/platform-build-control-plane/src/index.ts is 11,32
  • Claude_02 - Externalize Team Agent client defaults from the domain layer docs/Claude/Claude_02_externalize_client_defaults.md Claude 02 - Externalize Team Agent client defaults from the domain layer Priority: 2 Status: Revised after review Owner: TBD Last updated: 2026-05-22 Why this matters packages/domain/src/team-agent-work.ts is the domain core for Team Agent intake. It is otherw
  • Claude_03 — Decide what to do with `apps/platform` (and the untracked apps) docs/Claude/Claude_03_decide_apps_platform.md Claude 03 — Decide what to do with apps/platform and the untracked apps Priority: 3 Status: Proposed — needs a decision before any code moves Owner: TBD Last updated: 2026-05-22 Why this matters Three things in apps/ are misaligned with reality: 1. apps/platfo
  • Claude_04 - Validate environment configuration at Worker startup docs/Claude/Claude_04_env_validation_at_startup.md Claude 04 - Validate environment configuration at Worker startup Priority: 4 Status: Proposed Owner: TBD Last updated: 2026-05-22 Why This Matters The Env interface in apps/platform-build-control-plane/src/index.ts ../../apps/platform-build-control-plane/src/i
  • Claude_05 — Document the three-way skills split docs/Claude/Claude_05_document_skills_layers.md Claude 05 — Document the three-way skills split Priority: 5 lowest-risk, highest-clarity Status: Proposed Owner: TBD Last updated: 2026-05-22 Why this matters The word "skill" is used in three distinct ways in this repository, and nothing in the README or CONT
  • Claude_06 — Agent Runtime, Typed Tool Library, and Code-Executed Ad-hoc docs/Claude/Claude_06_tool_and_skill_registry_architecture.md Claude 06 — Agent Runtime, Typed Tool Library, and Code-Executed Ad-hoc Priority: Foundation Status: Proposed — second revision. Drops MCP as the internal tool protocol in favour of a typed TypeScript tool library + Cloudflare Sandbox for code execution. Owner
  • Client Data Plane Build Blueprint docs/client-data-plane-build-blueprint.md Client Data Plane Build Blueprint Status: draft implementation blueprint Updated: 2026-05-19 Current note: this blueprint documents the original fixture-first tracer bullet. ADR 0016 supersedes the live environment policy: current live client setup and runtime
  • Client Data Plane Specification docs/client-data-plane-specification.md Client Data Plane Specification Status: draft for review Updated: 2026-05-19 Purpose The Client Data Plane is the Cloudflare-first subsystem that stores client onboarding configuration, source bindings, ingestion runs, normalized facts, freshness state, and da
  • Client Performance Dashboard Architecture Proposal docs/client-facing-dashboard-architecture-proposal.html Client Performance Dashboard Architecture Proposal SUPERSEDED 2026-07-02 : Living Online has adopted a staff-only doctrine for Jessie and all Agency OS surfaces. No client-facing dashboard will be built from this proposal. It is retained for historical referen
  • Platform Build Loop Architecture Overview docs/platform-build-loop-architecture-overview.html Platform Build Loop Architecture Overview Living Online Agency OS Platform Build Loop architecture recap We are building the operating loop that lets Stuart and consultants manage the build of Agency OS from Asana, while Cloudflare coordinates the hosted contr
  • Plunkett SEO Reporting Agent Skill Proposal docs/plunkett-seo-reporting-agent-skill-proposal.html Plunkett SEO Reporting Agent Skill Proposal Agency OS Agent Skill Proposal Plunkett SEO reporting agent A first practical Agent Skill for one JWH brand and one channel: review SEO performance, explain what changed, recommend next actions, and write useful cont
  • AgentOS Box Data Source Runbook docs/runbooks/agentos-box-data-source.md AgentOS Box Data Source Runbook Status: v1 service-account source and Business Unit context gate scaffold Purpose AgentOS/Jessie uses Box as an approved client-context data source. The app authenticates as the AgentOS Box service account, not as a human user,
  • Box Agent Context Lake Operational Plan docs/runbooks/box-agent-context-lake-operational-plan.md Box Agent Context Lake Operational Plan Date: 2026-05-24 Status: Active implementation plan Executive Summary The current Box context implementation proved that Agency OS can bind a Business Unit to a Box Agent Context folder and version its file manifest. Tha
  • Runbook: Docker Desktop won't start — "Inference manager" crash-loop (blocks container deploys) docs/runbooks/docker-desktop-inference-crashloop.md Runbook: Docker Desktop won't start — "Inference manager" crash-loop blocks container deploys Impact: When this hits, the Docker daemon never comes up, so any Cloudflare container build/deploy fails — most importantly wrangler deploy of apps/execution-plane th
  • GA4 Reporting Skill Runbook docs/runbooks/ga4-reporting-skill.md GA4 Reporting Skill Runbook Purpose Generate an internal, review-required GA4 analytics report from a consultant Asana request. The report is designed for consultant scanning: major wins, major issues, and watch items appear at the top, followed by supporting
  • Google Ads Reporting Skill Runbook docs/runbooks/google-ads-reporting-skill.md Google Ads Reporting Skill Runbook Purpose The Google Ads reporting skill prepares an internal consultant-facing paid search report from approved context and a dated Google Ads data snapshot. The first slice is deterministic and review-required. It does not ch
  • Jessie Teams Conversation Slice Runbook docs/runbooks/jessie-teams-conversation-slice.md Jessie Teams Conversation Slice Runbook Status: v1 live reply scaffold Decision source: docs/adr/0007-jessie-teams-conversation-slice.md Purpose Jessie uses Microsoft Teams as a conversation and clarification surface. Teams is not the system of record for Deli
  • Local Agent Workspaces docs/runbooks/local-agent-workspaces.md Local Agent Workspaces Purpose This runbook keeps local Agency OS development clean when several agents and subagents are active at the same time. The root checkout at C:\Users\StuartWatson\Box\Living Online\agency-os is the coordinator workspace. It should st
  • Meta Ads Reporting Skill Runbook docs/runbooks/meta-ads-reporting-skill.md Meta Ads Reporting Skill Runbook Purpose Generate an internal, review-required Meta Ads report from a consultant Asana request. The first live client is Plunkett Homes. Example request: Required Configuration Cloudflare Worker vars: - TEAM AGENT META ADS DATA
  • Platform Build Loop Operator Runbook docs/runbooks/platform-build-loop-operator-runbook.md Platform Build Loop Operator Runbook Purpose This runbook describes how Builder Agents and human Admin Reviewers use the Platform Build Loop MVP to turn Asana Platform Build Tasks into checked implementation branches, Build Review Artifacts, and explicit human
  • Plunkett SEO Reporting Skill Runbook docs/runbooks/plunkett-seo-reporting-skill.md Plunkett SEO Reporting Skill Runbook Purpose The Plunkett SEO Reporting Skill is the first Agency OS client delivery agent component. It turns an Asana reporting request, an approved client context pack, and a dated SEO data snapshot into a consultant-facing r
  • Team Agent Client-Generic Runtime Runbook docs/runbooks/team-agent-client-generic-runtime.md Team Agent Client-Generic Runtime Runbook Status: implementation in progress Updated: 2026-05-24 Purpose This runbook defines how Jessie Team Agent execution moves from Plunkett-specific runtime wiring to client-generic execution driven by configured Business
  • Team Agent LLM and Client Context Runtime Plan docs/runbooks/team-agent-llm-context-runtime-plan.md Team Agent LLM and Client Context Runtime Plan Date: 2026-05-24 Status: Active remediation goal Goal Make Team Agent insight skills LLM-first and client-context-aware. Deterministic code may retrieve, normalize, and validate evidence, but it must not be the fi
  • Team Agent Work Intake Runbook docs/runbooks/team-agent-work-intake.md Team Agent Work Intake Runbook Purpose Team Agent Work is the first live path for staff to delegate delivery work through Asana. The local MVP lets an operator point the agent at an Asana task, verify that it is agent-eligible, post a Delegation Confirmation,
  • Technical SEO Audit Remediation Companion docs/runbooks/technical-seo-audit-remediation-companion.md Technical SEO Audit Remediation Companion Purpose This document defines the companion artifact that should be produced after a technical SEO audit has been reviewed and approved. The audit report explains what was found. The remediation companion translates th
  • Agent-Loop Execution Core Implementation Plan docs/superpowers/plans/2026-06-10-agent-loop-execution-core.md Agent-Loop Execution Core Implementation Plan For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development recommended or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox - syntax for tracking. Goal: R
  • Jessie Improvements Round 2 Implementation Plan docs/superpowers/plans/2026-06-12-jessie-improvements-round-2.md Jessie Improvements Round 2 Implementation Plan For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development recommended or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox - syntax for tracking. Goal:
  • Auto Context Management Implementation Plan docs/superpowers/plans/2026-06-14-auto-context-management.md Auto Context Management Implementation Plan For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development. Steps use checkbox - syntax. Model policy: Haiku/Sonnet implementers, lead reviews diffs. Goal: A scheduled job that reviews each
  • Goal: Complete ADR-0019 (Agent-Loop Execution Core) docs/superpowers/plans/2026-06-15-adr-0019-completion-goal.md Goal: Complete ADR-0019 Agent-Loop Execution Core Goal task prompt — paste into the Platform Build goal task; execute with subagents under lead direction. Goal statement Take ADR-0019 from "mostly built" to complete and operational : every skill runs on the pl
  • Harness Fallback Implementation Plan docs/superpowers/plans/2026-06-15-harness-fallback.md Harness Fallback Implementation Plan For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development recommended or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox - syntax for tracking. Goal: When a cla
  • Staff Admin Dashboard v1 Implementation Plan docs/superpowers/plans/2026-07-03-staff-admin-dashboard.md Staff Admin Dashboard v1 Implementation Plan For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development recommended or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox - syntax for tracking. Goal: A
  • Technical SEO Audit Reliability Plan (2026-07-06) docs/superpowers/plans/2026-07-06-technical-seo-reliability-plan.md Technical SEO Audit Reliability Plan 2026-07-06 Problem statement The technical-seo-audit skill cannot reliably produce a report on demand. Root cause is structural, not incidental: it is a ~30-minute batch job site crawl + analysis + report build + delivery e
  • Technical SEO: delivery now, execution plane earns its way back docs/superpowers/plans/2026-07-07-technical-seo-delivery-and-exec-plane-return.md Technical SEO: delivery now, execution plane earns its way back Date: 2026-07-07 Tuesday, ~2:00pm Perth Owner: Stuart Watson / session lead Supersedes nothing; extends docs/superpowers/plans/2026-07-06-technical-seo-reliability-plan.md with a hard separation o
  • Harvest to Dashboard Client Backfill Implementation Plan docs/superpowers/plans/2026-07-08-harvest-dashboard-client-backfill.md Harvest to Dashboard Client Backfill Implementation Plan For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development recommended or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox - syntax for tracki
  • Team Agent Meta-Ads Routing + Per-Client Meta Data Implementation Plan docs/superpowers/plans/2026-07-08-team-agent-meta-routing-fix.md Team Agent Meta-Ads Routing + Per-Client Meta Data Implementation Plan For agentic workers: This plan is executed by codex GPT-5.5 under lead review, per the repo's token-economy policy. Each task has precise anchors, a pattern to mirror, and acceptance tests.
  • Technical SEO Analyzer Upgrade: precision, evidence, and external-signal enrichment docs/superpowers/plans/2026-07-08-technical-seo-analyzer-upgrade.md Technical SEO Analyzer Upgrade: precision, evidence, and external-signal enrichment Date: 2026-07-08 Owner: Stuart Watson / session lead Status: proposed for review before implementation Why this exists The technical SEO report is only as good as the analysis
  • Jessie Channel Analysis And Recommendation Pipeline Implementation Plan docs/superpowers/plans/2026-07-10-channel-analysis-recommendation-pipeline.md Jessie Channel Analysis And Recommendation Pipeline Implementation Plan Date: 2026-07-10 Goal Make Jessie a dependable internal channel analyst for SEO including technical SEO , Google Ads, and Meta. A consultant assigns an Asana task to Jessie; Jessie reliabl
  • Bootstrap curation integration map (research, 2026-07-06) docs/superpowers/research/2026-07-06-bootstrap-curation-map.md Bootstrap curation integration map research, 2026-07-06 File:line verified against main on 2026-07-06 by a read-only research pass. Companion to docs/superpowers/specs/2026-07-06-context-bootstrap-curation-design.md. 1. Execution-plane bridge - Dispatch: index
  • Staff admin dashboard v1 - design (STAFF ONLY) docs/superpowers/specs/2026-07-03-staff-admin-dashboard-design.md Staff admin dashboard v1 - design STAFF ONLY Approved by Stuart 2026-07-03 mockup reviewed and accepted: "Yes, build it" . Purpose Give the admin team a one-glance, browser-accessible board of all configured clients: which platforms are configured for each, an
  • Box webhook context ingestion - design (STAFF ONLY) docs/superpowers/specs/2026-07-06-box-webhook-context-ingest-design.md Box webhook context ingestion - design STAFF ONLY Approved by Stuart 2026-07-06. Operating model decision Stuart : consultants own each client's Box "Agent Context" folder; when they drop a file in, Agency OS ingests it within about a minute. The daily cron sc
  • Autonomous bootstrap context curation - Phase 1 design (STAFF ONLY) docs/superpowers/specs/2026-07-06-context-bootstrap-curation-design.md Autonomous bootstrap context curation - Phase 1 design STAFF ONLY Approved by Stuart 2026-07-06: "manual is not the answer". Supersedes the consultants-copy-files bootstrap. Box content stays READ-ONLY ADR-0008 . Purpose When a client has an active Box binding
  • Harvest to Dashboard Client Backfill Sweep docs/superpowers/specs/2026-07-08-harvest-dashboard-client-backfill-design.md Harvest to Dashboard Client Backfill Sweep Date: 2026-07-08 Status: Design approved, pending implementation plan Problem Only a subset of the agency's active clients are configured in the admin dashboard. The dashboard reads the Admin Registry model in prod D1

Agent Skills

  • Jessie's voice skills/_shared/jessie-voice.md Jessie's voice You are Jessie — a sharp, warm marketing colleague at Living Online, not a reporting system. When you write anything a consultant reads Asana comments, report narrative, clarifying questions , follow these rules: - Lead with the so-what in plain
  • Aggregate Performance Reporting skills/aggregate-performance-reporting/SKILL.md --- name: aggregate-performance-reporting description: Build the Jessie Integrated Performance Report from SEO, GA4, Google Ads, and Meta evidence for consultant review before client delivery. --- Aggregate Performance Reporting Use this Jessie Team Agent skil
  • Context Pack Canonical Structure skills/client-onboarding/references/context-pack-template.md Context Pack Canonical Structure This file documents the canonical folder and file layout for a client context pack in Agency OS. Every new client context pack must follow this structure. All files must be present; unknown sections must carry a stub notice rat
  • Client Onboarding skills/client-onboarding/SKILL.md --- name: client-onboarding description: Onboard a new client into Agency OS through a normal Jessie task session. Use when an Asana task asks to add a new client, set up a client, register a client, configure client access, or create a client context pack. Pr
  • Context Curation skills/context-curation/SKILL.md --- name: context-curation description: Daily per-client Box context curation skill. Use when the scheduler asks Jessie to review a client's new or changed Box files, judge each candidate's marketing relevance, and produce a structured ingest/skip decision lis
  • GA4 Reporting skills/ga4-reporting/SKILL.md --- name: ga4-reporting description: Draft internal GA4 analytics reports from Asana task requests using live Google Analytics Data API evidence. Use when a task asks for a GA4, Google Analytics, website analytics, traffic, engagement, landing-page, conversion
  • Google Ads Reporting skills/google-ads-reporting/SKILL.md --- name: google-ads-reporting description: Agency OS Channel Skill for Google Ads reporting tasks. Use when an agent is asked to review paid search or Google Ads performance, explain spend/conversion movement, draft internal consultant reporting, recommend fo
  • Meta Ads Reporting skills/meta-ads-reporting/SKILL.md --- name: meta-ads-reporting description: Draft internal Meta Ads performance reports from Asana task requests using live Meta Marketing API evidence. Use when a task asks for a Meta Ads, Facebook Ads, paid social, or social advertising report, review, summary
  • Plunkett SEO Pilot skills/seo-reporting/references/plunkett-pilot.md Plunkett SEO Pilot Boundary The pilot proves one Agency OS Channel Skill for one client group, one business unit, and one channel: - Client group: JWH Group. - Business unit and pilot brand: Plunkett Homes / Plunkett. - Channel: SEO. - Primary task surface: As
  • SEO Reporting skills/seo-reporting/SKILL.md --- name: seo-reporting description: Agency OS Channel Skill for SEO reporting tasks. Use when an agent is asked to review SEO performance, explain organic search changes, draft consultant-facing reporting, recommend SEO follow-up work, write source-backed evi
  • Strategy Planning skills/strategy-planning/SKILL.md --- name: strategy-planning description: Build Jessie 90-day strategy plans from approved client context, active subscriptions, Harvest budget capacity, and consultant instructions. Use when an Asana task asks for a 90-day plan, strategy refresh, rolling quart
  • Technical SEO Audit skills/technical-seo-audit/SKILL.md --- name: technical-seo-audit description: Run a technical SEO audit from live crawl, PageSpeed, site-discovery, and rendered-validation evidence, then build the deterministic consultant report package through the technical seo analyze MCP tool. --- Technical