docs: update documentation to reflect 25 agents, 55 skills consolidation

Update README with accurate counts, install modules table, and CLI
reference. Remove premature release-notes directory. Refresh
ARCHITECTURE, authoring-guide, and workflow examples.

Author: jiten-singh-shahi

CHANGELOG.md

@@ -16,79 +16,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - `continuous-learning-v2` → `learning-engine` agent
   - `skill-stocktake` → merged into `sf-harness-optimizer` agent
 - Hook path `skills/continuous-learning-v2/hooks/observe.sh` → `scripts/hooks/learning-observe.sh`
-- Skills count: 65 → 59, Agents count: 27 → 32
-
-## [0.0.0] - 2026-03-26
-
-### Highlights
-
-- Initial release of Salesforce Claude Code (SCC) — production-ready AI agent harness for Salesforce development.
-- Built on the architecture and tooling of [Everything Claude Code (ECC)](https://github.com/affaan-m/everything-claude-code) by Affaan Mustafa, licensed under the MIT License.
-- 27 specialized agents covering Apex, LWC, SOQL, Flows, Visualforce, Aura, Agentforce, DevOps, security, and architecture.
-- 58 skills (39 user-invocable, 19 auto-activating) for Salesforce best practices, patterns, and workflows.
-- 28 rules across common, Apex, LWC, SOQL, Flow, Visualforce, and Aura domains.
-- 28 hooks across 7 lifecycle events with profile-based control (minimal, standard, strict).
-- Cross-harness support for Claude Code and Cursor with adapter-based content transformation.
-- AJV schema validation for state store, install state, install config, and CI validators.
-
-### Agents (27)
-
-- Salesforce specialists: sf-apex-reviewer, sf-lwc-reviewer, sf-trigger-architect, sf-soql-optimizer, sf-security-reviewer, sf-performance-optimizer, sf-flow-reviewer, sf-agentforce-builder, sf-data-architect, sf-integration-architect, sf-admin, sf-visualforce-reviewer, sf-aura-reviewer
-- Platform agents: sf-planner, sf-code-reviewer, sf-tdd-guide, sf-architect, sf-build-resolver, sf-docs-lookup, sf-devops-guide, sf-deployment-guide, sf-chief-of-staff, sf-harness-optimizer, sf-e2e-runner
-- Utility agents: doc-updater, refactor-cleaner, loop-operator
-
-### Skills (58)
-
-- 32 Salesforce-specific: sf-apex-best-practices, sf-apex-async-patterns, sf-apex-enterprise-patterns, sf-apex-testing, sf-apex-cursor, sf-lwc-development, sf-lwc-testing, sf-trigger-frameworks, sf-governor-limits, sf-soql-optimization, sf-integration, sf-deployment, sf-data-modeling, sf-debugging, sf-security, sf-devops-ci-cd, sf-scratch-org-workflow, sf-flow-development, sf-metadata-management, sf-metadata-migrations, sf-package-development, sf-agentforce-development, sf-api-design, sf-visualforce-development, sf-aura-development, sf-platform-events-cdc, sf-approval-processes, sf-experience-cloud, sf-reporting-dashboards, sf-docker-patterns, sf-e2e-testing, sf-tdd-workflow
-- 7 platform skills: configure-scc, continuous-agent-loop, mcp-server-patterns, prompt-optimizer, search-first, security-scan, strategic-compact
-- 13 workflow skills: sf-help, sf-quickstart, sf-build-fix, sf-harness-audit, checkpoint, sf-docs-lookup, refactor-clean, save-session, resume-session, sessions, update-docs, aside, model-route
-
-### Rules (28)
-
-- Common (11): agents, coding-style, development-workflow, git-workflow, governor-limits, hooks, naming-conventions, patterns, performance, security, testing
-- Apex (4): async, bulkification, coding-style, triggers
-- LWC (4): accessibility, coding-style, performance, testing
-- SOQL (2): optimization, security
-- Flow (3): best-practices, naming-conventions, testing
-- Visualforce (2): coding-style, security
-- Aura (2): coding-style, security
-
-### Hook System (28 hooks, 7 lifecycle events)
-
-- SessionStart: Salesforce project context detection (org info, CLI version, connected orgs)
-- PreToolUse: SF CLI validation, deprecated sfdx warnings, git push reminder, doc write warning, block-no-verify (protects git hooks)
-- PostToolUse: test coverage reminders (Apex/LWC), auto-format, typecheck, console.log detection, build/PR notifications
-- PostToolUseFailure: MCP health-check with exponential backoff and auto-reconnect
-- PreCompact: state preservation before context compaction
-- Stop: uncommitted changes reminder, session summary, cost tracking
-- SessionEnd: session metadata persistence, session evaluation
-- Quality gates: governor-check (SOQL/DML in loops), quality-gate (anti-pattern detection via apex-analysis), sfdx-scanner integration
-- Profile-based gating via `SCC_HOOK_PROFILE` (minimal, standard, strict) and `SCC_DISABLED_HOOKS`
-
-### Cross-Harness Support
-
-- Claude Code plugin manifest (`.claude-plugin/plugin.json`)
-- Cursor adapter layer: skill-adapter and agent-adapter transform content from Claude Code format to Cursor format
-- Build step (`npm run build:cursor`) generates `.cursor/skills/` (58) and `.cursor/agents/` (27) from source
-- Install-time transformation: `npx scc install --target cursor` auto-transforms during copy
-- Cursor hooks adapter (`.cursor/hooks/adapter.js`) delegates to shared `scripts/hooks/` implementations
-- MCP config auto-installed: `.mcp.json` (Claude Code) or `.cursor/mcp.json` (Cursor)
-
-### Schema Validation (8 schemas)
-
-- AJV-validated at runtime: state-store.schema.json (6 entity types), install-state.schema.json, scc-install-config.schema.json
-- AJV-validated in CI: hooks.schema.json, install-modules.schema.json, install-profiles.schema.json
-- Descriptive: plugin.schema.json, package-manager.schema.json
-- State store entity model: sessions, skillRuns, skillVersions, decisions, installState, governanceEvents
-- Graceful AJV fallback when not installed (bare environments)
-
-### Infrastructure
-
-- CLI tool (`npx scc`): install, uninstall, plan, doctor, repair, status, list-installed, sessions, session-inspect
-- Install config persistence: `--config scc-install.json` for team-shared install preferences
-- Selective installation via profiles (apex, lwc, full) and targets (claude, cursor)
-- Scripts organized by audience: `scripts/cli/` (end-user), `scripts/dev/` (contributor), `scripts/hooks/` (runtime), `scripts/lib/` (shared), `scripts/ci/` (validators)
-- npm package ships only what users need (cli + hooks + lib); dev/ci tools are repo-only
-- 8 CI validators for agents, commands, rules, skills, hooks, manifests, personal paths, catalog
-- 67 automated tests with 80% coverage threshold
-- GitHub Actions: CI (test + coverage), Validate (content structure), Lint (ESLint + markdownlint), Release (npm publish)
+- Consolidated to 25 agents and 55 skills (35 user-invocable, 20 auto-activating)
+- Removed phantom agents: sf-planner, sf-soql-optimizer, sf-chief-of-staff, sf-data-architect, sf-harness-optimizer, sf-docs-lookup (agent), sf-devops-guide, sf-deployment-guide
+- Removed phantom skills: sf-package-development, sf-docker-patterns, sf-metadata-migrations, sf-scratch-org-workflow, sf-reporting-dashboards
+- Upgraded CI validators: description limits 100-250 chars, SF keyword requirements, body structure checks, readonly/tools consistency
+- Aligned CI validators (validate-agents.js, validate-skills.js) with architect Python validators (validate_agent.py, validate_skill.py)
+- Renamed `### Constraints` → `### Guardrails` in 12 action skills to prevent constraint misclassification
+- Fixed install manifests to remove stale references and add missing constraint skills
+- Added release-please config for auto-bumping all plugin manifest versions
+- Removed premature docs/releases/ directory (release-please generates release notes)

CLAUDE.md

@@ -4,7 +4,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 ## Project Overview
 
-**Salesforce Claude Code (SCC)** is an agent harness performance system — a Claude Code plugin providing production-ready agents, skills, hooks, and MCP configurations specifically for Salesforce development. Published as `scc-universal` on npm (v1.0.0). Works across Claude Code and Cursor.
+**Salesforce Claude Code (SCC)** is an agent harness performance system — a Claude Code plugin providing production-ready agents, skills, hooks, and MCP configurations specifically for Salesforce development. Published as `scc-universal` on npm. Works across Claude Code and Cursor.
 
 This is the root directory for all SCC source code.
 

CONTRIBUTING.md

@@ -10,7 +10,6 @@ Thanks for wanting to contribute! This repo is a community resource for Salesfor
 - [Contributing Agents](#contributing-agents)
 - [Contributing Hooks](#contributing-hooks)
 - [Contributing User-Invocable Skills](#contributing-user-invocable-skills)
-- [Contributing Rules](#contributing-rules)
 - [Pull Request Process](#pull-request-process)
 
 ---
@@ -49,7 +48,7 @@ Useful Salesforce automations:
 
 Skills with `user-invocable: true` that users invoke via `/skill-name`:
 
-- Deployment skills (`/sf-deployment`, `/sf-scratch-org-workflow`)
+- Deployment skills (`/sf-deployment`, `/sf-devops-ci-cd`)
 - Testing skills (`/sf-tdd-workflow`, `/sf-apex-testing`, `/sf-e2e-testing`)
 - Security skills (`/sf-security`, `/sf-governor-limits`)
 - Platform skills (`/continuous-agent-loop`, `/prompt-optimizer`, `/strategic-compact`)
@@ -97,7 +96,7 @@ skills/
 ```markdown
 ---
 name: your-skill-name
-description: Brief description shown in skill list
+description: "Use when [trigger] for Salesforce [domain]. Do NOT use for [exclusions]. (100-250 chars, 3+ SF keywords)"
 origin: SCC
 ---
 
@@ -149,9 +148,10 @@ agents/your-agent-name.md
 ```markdown
 ---
 name: your-agent-name
-description: What this agent does and when Claude should invoke it.
+description: "Use when [trigger]. Do NOT use for [exclusions]. 100-250 chars with 3+ Salesforce keywords."
 tools: ["Read", "Write", "Edit", "Bash", "Grep", "Glob"]
 model: sonnet
+origin: SCC
 ---
 
 You are a [role] specialist for Salesforce development.
@@ -185,7 +185,8 @@ What you return to the user.
 | `name` | Lowercase, hyphenated | `sf-apex-reviewer` |
 | `description` | Used to decide when to invoke | Be specific! |
 | `tools` | Only what's needed | `Read, Write, Edit, Bash, Grep, Glob, WebFetch, Task` |
-| `model` | Complexity level | `haiku` (simple), `sonnet` (coding), `opus` (complex) |
+| `model` | Complexity level | `haiku` (simple), `sonnet` (coding), `opus` (complex), `inherit` (caller decides) |
+| `origin` | Must be SCC | `SCC` |
 
 ---
 
@@ -247,7 +248,7 @@ skills/your-skill-name/SKILL.md
 ```markdown
 ---
 name: your-skill-name
-description: Brief description of what this skill does (minimum 30 characters)
+description: "Use when [trigger] for Salesforce [domain]. Do NOT use for [exclusions]. (100-250 chars, 3+ SF keywords)"
 origin: SCC
 user-invocable: true
 ---
@@ -293,7 +294,6 @@ What you're adding and why.
 - [ ] Skill
 - [ ] Agent
 - [ ] Hook
-- [ ] Rule
 
 ## Testing
 How you tested this with Salesforce projects.