The Ultimate Claude Code Skills Playbook — Solanasis LLC

Date: 2026-03-23 Version: 1.0 Status: AUTHORITATIVE — This is the single source of truth for all Claude Code skills, plugins, and capability decisions. Author: Claude Code (Opus 4.6), commissioned by Dmitri Zasage Research basis: 6 deep-research agents, 40+ repositories evaluated, 80+ official plugins catalogued, 12 business functions mapped

Supersession Table

This playbook supersedes these prior documents. Each should be archived to archive/2026-03/ with a header note.

Prior Document	What It Covered	Status
`claude-skills-acquisition-plan.md`	Skill evaluation across 3 tiers (685 lines)	Merged into Sections 3 + 7
`solanasis-cowork-master-setup-plan.md`	MCP servers, tool routing, CLAUDE.md config (750 lines)	Merged into Sections 1 + 2
`INSTALLATION-PLAN-MCP-PLUGINS-SKILLS.md`	Step-by-step installation guide	Merged into Section 5
`INSTALLATION-QUICK-CHECKLIST.md`	Action checklist	Merged into Section 5

Keep as separate reference: claude-code_gsd_codex_model-council_playbook_handoff_2026-03-19.md (multi-model orchestration — different topic)

Current State (Verified 2026-03-23)

Component	Status
`~/.claude/skills/`	Created, empty
Project `skills/`	Exists, empty
Installed plugins	Only `frontend-design` (2 copies)
Marketplaces	`claude-plugins-official` (80+), `claude-code-plugins` (14 demo)
`knowledge-work-plugins`	NOT added — 13 business plugins inaccessible
Subagents	3 working: planner, research-agent, senior-reviewer
MCP servers (local)	None in settings.json
MCP connectors (Claude.ai)	ClickUp, Gmail, Google Calendar, Canva, Fireflies, Gamma
CLAUDE.md	Configured (user + project level)
Permissions	bypassPermissions mode active

1. Architecture — How Skills Work in Claude Code

The Three Layers

CLAUDE.md (always loaded) — Persistent instructions guiding behavior across sessions
Subagents (~/.claude/agents/) — Specialized agents for planning, research, review
Skills (~/.claude/skills/ or .claude/skills/) — Modular capabilities in SKILL.md format

Skills vs Plugins vs Commands

Concept	What It Is	Scope
Skill	A SKILL.md file with instructions + optional scripts/resources. Claude loads it when relevant or you invoke with `/skill-name`.	Single capability
Plugin	A package bundling multiple skills, commands, agents, hooks, and MCP servers. Has `.claude-plugin/plugin.json`.	Distributable bundle
Command	Legacy term — `.claude/commands/` files still work but are now treated as skills.	Single action

SKILL.md Format

---
name: skill-name              # Required. Max 64 chars, lowercase + hyphens
description: What this does    # Required. Max 1024 chars. Include trigger keywords.
disable-model-invocation: false # true = manual-only (saves context budget)
context: fork                   # Optional. Runs as subagent (isolates context)
---
 
# Instructions here (markdown)

Context Budget

Budget: 2% of context window = ~16,000 characters for skill descriptions
Per skill: ~250-360 chars (description + XML overhead)
Practical limit: 40-50 model-invocable skills before truncation
Recommended: 12-15 auto-invoke for quality trigger accuracy
Override: Set SLASH_COMMAND_TOOL_CHAR_BUDGET env var for custom limit
Check: Run /context to see “Excluded skills” warnings

Progressive Disclosure

Metadata (~100 tokens): name + description loaded at startup for ALL skills
Instructions (<5000 tokens): Full SKILL.md body loaded only when skill activates
Resources (on demand): scripts/, references/, assets/ loaded when needed

Pro tip: Keep SKILL.md under 500 lines. Move reference material to references/ files. One level deep only — no SKILL.md → file1 → file2 chains.

Agent Skills Open Standard

Skills built for Claude Code also work in: OpenAI Codex CLI, GitHub Copilot, VS Code, Gemini CLI, Cursor, Windsurf, Aider, Kiro (AWS), and 20+ other platforms. Spec at agentskills.io.

Implication for Solanasis: Custom skills you build are portable. They’re not locked to Claude. This reduces vendor risk and could become deliverables for clients.

2. The Three-Marketplace Strategy

Marketplace 1: `claude-plugins-official` (Auto-available)

Already on your machine. 80+ plugins. Run /plugin → Discover tab.

Top plugins by install count:

Plugin	Installs	What It Does
`frontend-design`	371K	Production-grade UI (INSTALLED)
`superpowers`	233K	Brainstorming, subagent dev, TDD, code review
`context7`	189K	Version-specific documentation lookup
`code-review`	169K	5 parallel review agents with confidence scoring
`code-simplifier`	140K	Code clarity improvements
`feature-dev`	131K	7-phase dev with explorer/architect/reviewer agents
`playwright`	118K	Browser automation and E2E testing

Solanasis-relevant plugins:

commit-commands — /commit, /commit-push-pr (daily git workflow)
security-guidance — PreToolUse hook monitoring 9 security patterns
hookify — Custom behavior hooks
skill-creator — Create, test, benchmark skills
searchfit-seo — SEO toolkit
zapier — 8,000+ app integrations
zoominfo — B2B sales enrichment
aikido — SAST, secrets, IaC vulnerabilities
semgrep — Real-time security vulnerability detection
LSP plugins: pyright-lsp, typescript-lsp, csharp-lsp (for active dev languages)

Marketplace 2: `knowledge-work-plugins` (Must Add Manually)

Install command: claude plugin marketplace add anthropics/knowledge-work-plugins

13 role-based plugins, 10.3K GitHub stars, Apache 2.0 license. Confirmed compatible with Claude Code CLI.

Plugin	Skills	Commands	Key Connectors
Sales	6	3	HubSpot, Close, Clay, ZoomInfo, Fireflies, Gong
Operations	6	6	ServiceNow, Zendesk, Jira, ClickUp, Notion, Confluence
Engineering	6	6	GitHub, GitLab, Datadog, PagerDuty
Marketing	5	7	Canva, HubSpot, Ahrefs
Legal	6	5	Box, Jira
Customer Support	5	5	Intercom, HubSpot, Guru
Data	6	6	Snowflake, Databricks, BigQuery
Human Resources	6	6	Workday, BambooHR, Greenhouse
Finance	6	5	Snowflake, Databricks
Product Management	7	7	Linear, ClickUp, Figma, Amplitude
Enterprise Search	3	2	Slack, email, cloud storage
Design	6	6	Figma, Amplitude
Productivity	2	3	ClickUp, Notion, Asana, Linear, Jira

Bold = matches Solanasis current stack (ClickUp, Canva, Fireflies).

Marketplace 3: `claude-code-plugins` (Demo/Examples)

14 plugins from anthropics/claude-code. Already added. Most overlap with claude-plugins-official. Useful demo: commit-commands, hookify, plugin-dev.

3. The Solanasis Skills Stack — By Business Function

3.1 — Sales & Pipeline

Official (install Week 1):

account-research — Company overview, recent news, key people, tech stack (AUTO-INVOKE)
call-prep — Prep brief with attendee bios, agenda, prior relationship context (AUTO-INVOKE)
draft-outreach — Research-first outreach: email + LinkedIn message (AUTO-INVOKE)
create-an-asset — Landing pages, decks, one-pagers, workflow demos (manual)
competitive-intelligence — Interactive HTML battlecard (manual)
daily-briefing — Prioritized daily brief from calendar, pipeline, email (manual)

Community (supplement in Month 2):

ai-sales-team-claude (zubair-trabzada) — 14 skills, 5 parallel agents. BANT+MEDDIC qualification, proposal generation, PDF pipeline reports
gtm-flywheel (kenny589) — 15 skills. SPARK cold email framework, ICP scoring, signal taxonomy

Custom build needed:

Fractional CIO battlecards (fCIO vs full-time vs MSP vs freelancer)

3.2 — Marketing & Content

Official (install Week 2):

content-creation — Blog, social, email, landing page, case study templates (AUTO-INVOKE)
campaign-planning — OAMCM framework, content calendar, channel strategy (manual)
brand-voice — Brand enforcement with tone-by-context matrix (manual)
competitive-analysis — Positioning analysis, content gap detection (manual)
performance-analytics — Channel metrics with benchmark ranges (manual)

Community (install Week 2):

claude-blog (AgriciDaniel) — 19 sub-skills, 12 content templates, Astro CMS support, dual Google+AI citation optimization, fact-checking pipeline, 100-point quality scoring
cold-email + pricing-strategy (coreyhaines31/marketingskills) — Campaign-tested cold email, pricing decisions for consulting

Custom build needed:

LinkedIn post drafter (trained on Solanasis voice profile)
Brevo email campaign skill (List ID: 2)
fCIO case study template

3.3 — SEO & AI Search

Community (install Week 2):

claude-seo (AgriciDaniel) — 13 sub-skills, 7 subagents. Technical SEO, E-E-A-T, schema, Core Web Vitals (INP), GEO/AEO for AI search, programmatic SEO, DataForSEO MCP (22 commands). Strategic planning with agency model.
gtm-engineer-skills (onvoyage-ai) — Specialized GEO/AEO. 16 checks, 80+ AEO score target. Critical for AI search visibility: “Who are the best fractional CIOs in Colorado?”

Official marketplace: searchfit-seo@claude-plugins-official

3.4 — Legal & Compliance

Official (install Week 2):

contract-review — Clause-by-clause with GREEN/YELLOW/RED flags (AUTO-INVOKE)
nda-triage — Three-bucket categorization (manual)
compliance — GDPR, CCPA, HIPAA, LGPD (manual)
legal-risk-assessment — Severity x likelihood, 1-25 scoring (manual)
canned-responses — Templates for DSR, discovery holds (manual)

Community (install Week 1-2):

GRC Skills (Sushegaad) — 99% eval accuracy. 5 skills:
- ISO 27001: Lead Auditor, 2013 (114 controls) + 2022 (93 controls), gap assessments, SoA templates (AUTO-INVOKE)
- SOC 2: AICPA 2017 TSC, traffic-light gap analysis, 12 policy templates, 32-question vendor risk assessment (AUTO-INVOKE)
- HIPAA: Privacy/Security/Breach Notification Rules, 9 document templates
- GDPR: Dual EU/UK, DPA/DPIA/Privacy Notice templates
- FedRAMP: 75+ item readiness checklist, NIST SP 800-53 Rev 5 mapping
claude-legal-skill (evolsb) — CUAD grounded (41 risk categories, 510 real contracts). Position-aware review. Market benchmarks. Tracked-changes Word + PDF redlines. Jurisdiction intelligence.
OWASP skill (agamm) — OWASP Top 10:2025, ASVS 5.0, Agentic AI Security

Custom build needed:

Solanasis SOW/MSA template (50/50 terms, scope, liability)
Compliance framework crosswalk (SOC 2 ↔ ISO 27001 ↔ NIST CSF ↔ HIPAA)
Third-party risk management questionnaire generator

3.5 — Security Assessment & DR (Your Wedge)

Community (install Month 2):

Trail of Bits (trailofbits/skills) — 45 skills, industry gold standard. Key:
- insecure-defaults — Hardcoded creds, fail-open patterns
- supply-chain-risk-auditor — Dependency threat landscape
- static-analysis — CodeQL, Semgrep, SARIF
- audit-context-building — Deep architecture analysis
- differential-review — Security-focused code change review
- sharp-edges — Error-prone APIs, dangerous configs
Trail of Bits config (trailofbits/claude-code-config) — Settings templates, hook examples

Official marketplace:

security-guidance@claude-plugins-official — PreToolUse hook, 9 security patterns
aikido@claude-plugins-official — SAST, secrets, IaC
semgrep@claude-plugins-official — Vulnerability detection

Custom build needed (CRITICAL — core offerings):

security-assessment-report — .docx/.pdf with executive summary, findings (critical/high/medium/low), NIST CSF/CIS Controls mapping, remediation roadmap, compliance matrix
dr-verification-workflow — Backup restoration, failover testing, communication plan, RTO/RPO measurement, pass/fail report
incident-response-playbook-generator — SMB-tailored IR playbooks

3.6 — Engineering & Development

Official (install Week 1):

documentation — README, API docs, runbooks, architecture docs, onboarding (AUTO-INVOKE)
incident-response — SEV1-4, 5-step framework, blameless postmortem (AUTO-INVOKE)
code-review — 4-dimension: security, performance, correctness, maintainability (manual)
system-design — 5-step architecture framework (manual)
tech-debt — 6 debt categories with scoring (manual)
testing-strategy — Testing pyramid guidance (manual)

Official marketplace (install Phase 1-2):

superpowers@claude-plugins-official — Brainstorming, TDD, subagent dev
context7@claude-plugins-official — Version-specific docs
commit-commands@claude-plugins-official — Git workflow
playwright@claude-plugins-official — E2E testing for solanasis.com

3.7 — Operations

Official (install Week 1):

vendor-management — TCO analysis, risk assessment, performance metrics (AUTO-INVOKE)
process-optimization — Lean-style 4-phase improvement (AUTO-INVOKE)
risk-assessment — Likelihood x Impact, 6 risk categories (AUTO-INVOKE)
compliance-tracking — SOC 2, ISO 27001, GDPR, HIPAA, PCI DSS (AUTO-INVOKE)
change-management — 4-phase framework (manual)
resource-planning — People, budget, time allocation (manual)

Community:

SOP Structure (TheBushidoCollective) — SOP anatomy for contractor SOPs
pm-skills (phuryn) — 100+ skills: OKRs, roadmaps, sprints, retrospectives

3.8 — Finance

Official (install when needed):

Finance plugin (6 skills) — journal entries, reconciliation, financial statements, variance analysis, audit-support

Community:

Composio Xero MCP — Direct Xero API: invoices, transactions, reports
financial-analyst (alirezarezvani) — Ratio analysis, DCF, forecasts

3.9 — HR

Official (install when scaling):

HR plugin (6 skills) — recruiting, onboarding, compensation, org planning

Community:

hr-skills (tuanductran) — 15+ HR lifecycle skills

Custom: 1099 contractor onboarding skill

3.10 — Executive & Strategic

Community (install Month 2):

C-Level Advisory (alirezarezvani) — 28 skills. CEO, CTO, CFO, CMO, COO advisor roles with strategic frameworks. Directly mirrors fCIO/fCSIO/fCOO positioning.
Skill Factory (alirezarezvani/claude-code-skill-factory) — 5 builder agents, 10 slash commands for building custom skills

Custom: Strategic tech roadmap generator, QBR deck generator

3.11 — Customer Success

Official (install when client count reaches 3+):

Customer Support plugin (5 skills) — triage, research, response, escalation, knowledge management

3.12 — Data & Analytics

Official (install when migration scoped):

Data plugin (6 skills) — SQL, exploration, visualization, validation, dashboards

Community:

data-engineering-skills (AltimateAI) — +22% faster queries, migration architecture

Custom: CRM data migration playbook, SMB data quality assessment

4. Context Window Budget Plan

Auto-Invoke Roster (15 skills)

#	Skill	Source	Est. Chars
1	account-research	Sales plugin	~300
2	call-prep	Sales plugin	~300
3	draft-outreach	Sales plugin	~300
4	compliance-tracking	Operations plugin	~300
5	risk-assessment	Operations plugin	~300
6	vendor-management	Operations plugin	~300
7	process-optimization	Operations plugin	~300
8	documentation	Engineering plugin	~300
9	incident-response	Engineering plugin	~300
10	content-creation	Marketing plugin	~300
11	frontend-design	Official (installed)	already loaded
12	contract-review	Legal plugin	~300
13	SOC 2 GRC	Sushegaad	~300
14	ISO 27001 GRC	Sushegaad	~300
15	security-assessment-report	Custom build	~300

Total: ~4,200 of ~16,000 chars (26%). Safe margin for growth.

Manual-Only (invoke with `/skill-name`)

Everything else. Heavy skills (claude-seo, Trail of Bits, ai-sales-team) should use context: fork to run as subagents.

Monitoring

Weekly: Run /context — look for “Excluded skills” warnings
Monthly: Review which auto-invoke skills fired correctly vs incorrectly
Quarterly: Prune unused skills, rebalance budget

5. Installation Sequence

Phase 0: Prerequisites (30 min)

# 1. Skills directory already created at ~/.claude/skills/
 
# 2. Add knowledge-work-plugins marketplace
claude plugin marketplace add anthropics/knowledge-work-plugins
 
# 3. Verify
# Run /plugin → Discover tab should show Sales, Operations, Marketing, etc.

Phase 1: Revenue Engine — Week 1 (6-8 hours)

# 4. Sales plugin (6 skills)
claude plugin install sales@knowledge-work-plugins
# Set manual-only: create-an-asset, competitive-intelligence, daily-briefing
 
# 5. Operations plugin (6 skills)
claude plugin install operations@knowledge-work-plugins
# Set manual-only: change-management, resource-planning
 
# 6. Engineering plugin (6 skills)
claude plugin install engineering@knowledge-work-plugins
# Set manual-only: code-review, system-design, tech-debt, testing-strategy
 
# 7. Development workflow plugins
claude plugin install commit-commands@claude-plugins-official
claude plugin install superpowers@claude-plugins-official

Then manually install GRC skills:

Clone Sushegaad/Claude-Skills-Governance-Risk-and-Compliance
Copy ISO 27001 and SOC 2 skill directories to ~/.claude/skills/

Verify: Test account-research on a real prospect. Test compliance-tracking with “SOC 2 gap analysis.”

Phase 2: Marketing & Legal — Week 2 (6-8 hours)

# 11. Marketing plugin
claude plugin install marketing@knowledge-work-plugins
# Set manual-only: campaign-planning, brand-voice, competitive-analysis, performance-analytics
 
# 12. Legal plugin
claude plugin install legal@knowledge-work-plugins
# Set manual-only: all except contract-review

Then manually install community skills:

Clone AgriciDaniel/claude-blog → copy to .claude/skills/claude-blog/
Clone AgriciDaniel/claude-seo → copy to .claude/skills/claude-seo/
Clone evolsb/claude-legal-skill → copy to ~/.claude/skills/claude-legal/

Verify: Test content-creation on a blog post. Test contract-review on template NDA.

Phase 3: Custom Builds — Weeks 3-4 (8-10 hours)

Using skill-creator:

Build security-assessment-report — NIST CSF + CIS Controls mapping, .docx output
Build client-onboarding — Discovery questionnaire → systems inventory → kickoff
Build dr-verification-workflow — Test plan → execution → gap analysis → report
Add remaining GRC skills (HIPAA, GDPR, FedRAMP) to ~/.claude/skills/

Verify: Generate sample security assessment → run senior-reviewer audit. Context budget check: /context — no warnings.

Phase 4: Depth — Month 2 (ongoing)

# As needed:
claude plugin install data@knowledge-work-plugins
claude plugin install customer-support@knowledge-work-plugins
claude plugin install aikido@claude-plugins-official
claude plugin install semgrep@claude-plugins-official

Plus:

Clone Trail of Bits skills (cherry-pick: insecure-defaults, supply-chain-risk-auditor, static-analysis)
Cherry-pick alirezarezvani C-Level Advisory skills
Clone ai-sales-team-claude for BANT/MEDDIC methodology
Evaluate Wyre Technology MSP platform

Phase 5: Optimization — Month 3+ (quarterly)

Prune unused skills
Update descriptions for better trigger accuracy
Context budget rebalance
Evaluate new marketplace plugins (ecosystem growing weekly)
Build remaining custom skills as pipeline demands

6. Building Custom Skills

Solanasis Template

---
name: skill-name
description: What this skill does and when Claude should use it. Be specific.
  Include trigger keywords relevant to Solanasis operations.
disable-model-invocation: false
---
 
# Skill Name
 
## When to Use
- [Trigger condition 1]
- [Trigger condition 2]
 
## Process
1. [Step 1]
2. [Step 2]
3. [Step 3]
 
## Output Format
[Expected deliverable structure]
 
## Quality Checklist
- [ ] [Quality gate 1]
- [ ] [Quality gate 2]
 
## References
- See `references/` directory for supporting materials

Priority Custom Skills

#	Skill	What It Generates	Build When
1	security-assessment-report	Client-ready .docx: exec summary, findings matrix, NIST/CIS mapping, remediation roadmap	Week 3
2	client-onboarding	Discovery questionnaire, systems inventory, stakeholder map, kickoff package	Week 3
3	dr-verification-workflow	DR test plan, execution log, gap analysis, pass/fail report	Week 4
4	solanasis-sow-template	SOW with 50/50 terms, scope clauses, liability limitations	Week 4
5	data-migration-planner	7-phase migration runbook (discovery → validation → cutover)	When first migration scoped
6	compliance-crosswalk	SOC 2 ↔ ISO 27001 ↔ NIST CSF ↔ HIPAA control mapping	Month 2
7	responsible-ai-assessment	NIST AI RMF + ISO 42001 assessment report	Month 2
8	strategic-tech-roadmap	12-24 month technology plans for SMBs	Month 2

Build Process

Use skill-creator (from claude-plugins-official) for scaffolding
Write SKILL.md following the template above
Add reference files to references/ subdirectory
Test with 3+ real-world scenarios
Run senior-reviewer on generated output
Iterate descriptions until trigger accuracy is reliable
Set disable-model-invocation appropriately

7. What to Skip (and Why)

Skip	Reason
antigravity-awesome-skills (1,304)	Quantity over quality. Use as discovery portal only.
jeremylongshore collection (1,367)	Same — noise, not signal.
SEO Machine (TheCraigHewitt)	Overlaps with claude-seo. WordPress-centric (you use Astro).
Bio-Research plugin	Not relevant.
Design plugin	Canva MCP covers this.
ralph-loop / ralph-wiggum	Autonomous loops without human check. Dangerous.
claude-opus-4-5-migration	Already on Opus 4.6.
learning/explanatory output styles	AI-native agency = autonomous work, not teaching.
OneWave-AI/claude-skills	19 sports skills in “production” collection = padding.
Most LSP plugins	Only install for active languages (Python, TS, C#).
Finance plugin (initially)	Xero handles current needs.
HR plugin (initially)	One-person operation.
Product Management plugin	Service business, not product business.

8. Verification & Monitoring

Post-Installation Tests (Run After Each Phase)

/plugin → confirm all installed plugins appear
Research a real prospect with account-research → structured brief
Draft outreach with draft-outreach → personalized email with research hooks
Run compliance-tracking → SOC 2 gap analysis output
Run contract-review on template NDA → GREEN/YELLOW/RED flags
Run content-creation for a blog post → SEO-structured draft
Generate security assessment with custom skill → .docx with NIST mapping
Generate client onboarding package → discovery questionnaire + kickoff
Run ISO 27001 GRC skill → gap assessment with SoA template
Check /context → no “Excluded skills” warnings

Weekly Monitoring (5 min)

Check /context for truncation warnings
Note which skills auto-triggered correctly/incorrectly
Adjust descriptions for misfiring skills

Quarterly Review

Prune skills not used in 30+ days
Check marketplace for new relevant plugins
Rebalance context budget allocation
Run senior-reviewer on all custom skill outputs
Update this playbook’s Current State table

9. Marketplace & Discovery Reference

Official Anthropic Sources

Source	Stars	What
anthropics/skills	101K	Document skills, skill-creator, Agent Skills spec
anthropics/knowledge-work-plugins	10.3K	13 business plugins (Sales, Ops, Legal, etc.)
anthropics/claude-code	81.8K	Demo plugins
anthropics/claude-plugins-official	14.3K	80+ official marketplace
agentskills.io	—	Open standard specification

Top Community Sources

Source	Stars	What	Solanasis Relevance
Sushegaad/GRC	62	ISO 27001, SOC 2, HIPAA, GDPR, FedRAMP (99% eval)	CRITICAL
trailofbits/skills	3.8K	45 security audit skills	HIGH
alirezarezvani/claude-skills	6.6K	205 skills, C-level advisory	HIGH
coreyhaines31/marketingskills	15.9K	34 growth marketing skills	HIGH
AgriciDaniel/claude-seo	3K	13 SEO sub-skills, DataForSEO	HIGH
AgriciDaniel/claude-blog	—	19 blog skills, Astro support	HIGH
evolsb/claude-legal-skill	115	CUAD contract review, redlines	HIGH
wyre-technology/msp-claude-plugins	3	MSP platform integrations	EVALUATE

Aggregators (Discovery, Not Bulk Install)

VoltAgent/awesome-agent-skills — Official vendor skills directory
ComposioHQ/awesome-claude-skills — 500+ app integrations
sickn33/antigravity-awesome-skills — 1,304+ browsable skills
SkillsMP.com — 500K+ indexed skills
MCPMarket.com — Skills and MCP servers

Evaluation Criteria

Before installing any community skill:

Read the SKILL.md source (not just the README)
Check maintenance: star count, last commit, contributor count
Verify license: Apache 2.0 or MIT preferred
Test with 2+ scenarios before adding to auto-invoke
Check context cost: description length affects budget

10. Prior Document Archive

After this playbook is finalized, archive these to solanasis-docs/archive/2026-03/:

Document	Archive Header
`claude-skills-acquisition-plan.md`	`> ARCHIVED 2026-03-23 — Superseded by playbooks/ultimate-claude-code-skills-playbook.md`
`solanasis-cowork-master-setup-plan.md`	Same
`INSTALLATION-PLAN-MCP-PLUGINS-SKILLS.md`	Same
`INSTALLATION-QUICK-CHECKLIST.md`	Same

Keep as reference: claude-code_gsd_codex_model-council_playbook_handoff_2026-03-19.md

This playbook is a living document. Update Section “Current State” after each installation phase. This is the single source of truth for Claude Code capabilities at Solanasis.

Solanasis Docs

Explorer

ultimate-claude-code-skills-playbook

The Ultimate Claude Code Skills Playbook — Solanasis LLC

Supersession Table

Current State (Verified 2026-03-23)

1. Architecture — How Skills Work in Claude Code

The Three Layers

Skills vs Plugins vs Commands

SKILL.md Format

Context Budget

Progressive Disclosure

Agent Skills Open Standard

2. The Three-Marketplace Strategy

Marketplace 1: claude-plugins-official (Auto-available)

Marketplace 2: knowledge-work-plugins (Must Add Manually)

Marketplace 3: claude-code-plugins (Demo/Examples)

3. The Solanasis Skills Stack — By Business Function

3.1 — Sales & Pipeline

3.2 — Marketing & Content

3.3 — SEO & AI Search

3.4 — Legal & Compliance

3.5 — Security Assessment & DR (Your Wedge)

3.6 — Engineering & Development

3.7 — Operations

3.8 — Finance

3.9 — HR

3.10 — Executive & Strategic

3.11 — Customer Success

3.12 — Data & Analytics

4. Context Window Budget Plan

Auto-Invoke Roster (15 skills)

Manual-Only (invoke with /skill-name)

Monitoring

5. Installation Sequence

Phase 0: Prerequisites (30 min)

Phase 1: Revenue Engine — Week 1 (6-8 hours)

Phase 2: Marketing & Legal — Week 2 (6-8 hours)

Phase 3: Custom Builds — Weeks 3-4 (8-10 hours)

Phase 4: Depth — Month 2 (ongoing)

Phase 5: Optimization — Month 3+ (quarterly)

6. Building Custom Skills

Solanasis Template

Priority Custom Skills

Build Process

7. What to Skip (and Why)

8. Verification & Monitoring

Post-Installation Tests (Run After Each Phase)

Weekly Monitoring (5 min)

Quarterly Review

9. Marketplace & Discovery Reference

Official Anthropic Sources

Top Community Sources

Aggregators (Discovery, Not Bulk Install)

Evaluation Criteria

10. Prior Document Archive

Graph View

Table of Contents

Backlinks

Marketplace 1: `claude-plugins-official` (Auto-available)

Marketplace 2: `knowledge-work-plugins` (Must Add Manually)

Marketplace 3: `claude-code-plugins` (Demo/Examples)

Manual-Only (invoke with `/skill-name`)