Test

Install, fully integrate, and prove all 15 skills below. Do not just clone — configure, wire each into the agent workflow, and demonstrate end-to-end that it works. Work autonomously; never stop to ask permission; come back only with results or a genuine blocker.Before starting:- Verify every URL actually exists. If a repo 404s, is empty, or has no install docs — report it plainly and continue. Never invent an install for a repo that isn’t there.- Read each repo’s full README + docs before touching it.- Map the agent’s skill system: SKILL.md format, skills folder, MCP config location, rules file.The 15:1. SkillClaw — https://github.com/AMAP-ML/SkillClaw2. skills — https://github.com/mattpocock/skills3. defuddle — https://github.com/kepano/defuddle4. humanizer — https://github.com/blader/humanizer5. youtube-skills — https://github.com/ZeroPointRepo/youtube-skills6. composio — https://github.com/ComposioHQ/composio7. agent-skills — https://github.com/addyosmani/agent-skills8. detect-skill — https://github.com/resemble-ai/detect-skill9. OpenMontage — https://github.com/calesthio/OpenMontage10. Anthropic-Cybersecurity-Skills — https://github.com/mukul975/Anthropic-Cybersecurity-Skills11. oh-my-hermes — https://github.com/rlaope/oh-my-hermes12. browser-harness — https://github.com/browser-use/browser-harness13. codebase-memory-mcp — https://github.com/DeusData/codebase-memory-mcp14. loopy — https://github.com/Forward-Future/loopy15. Agent-Reach — https://github.com/Panniantong/Agent-ReachFor each: read its README, install exactly per its official instructions, and register it as a callable skill in the agent’s skill system. If the repo is MCP-based (composio, codebase-memory-mcp, browser-harness likely are), wire it into the MCP config (cline_mcp_settings.json or equivalent), not just disk.Full integration — mandatory for all 15:- Add a rule to the rules file (AGENTS.md / .clinerules) so each skill auto-fires when relevant — the agent must not need to be told to use it.- Write a one-line trigger per skill: exactly when the agent should invoke it.- Any skill needing config tweaks to work in this environment: make them and document why.Final verification — all 15, no exceptions:- List where each skill lives and how it’s invoked.- Prove each one with a real end-to-end test (live output, generated file, or screenshot — not a claim).- One broken skill = the whole job is unfinished. Keep working until all 15 pass.- Log every command: command → output → success/fail. Show the receipts.Commandments:1. No “Done!” without proof — every install ends in a visible artifact.2. Read the skill before declaring it installed; be able to summarize what it does.3. One broken skill = whole job failed. No “almost everything works.”4. No interactive prompts — find the non-interactive flag or CI mode, never hang.5. Docs conflict? Official README wins. Ambiguous? Test both ways, keep what works.6. Wire it, don’t store it — skills that need manual invocation are dead weight.7. Log every command. I want the receipts.8. Never stop to ask permission. Decide, execute, fix, report.9. Test in the real environment — no dry runs, no mocks.10. A 404 or dead repo is a finding, not a failure — report it and move on.11. If two skills conflict (same folder, same command name, same MCP port), resolve it and document the resolution.12. Leave it better than you found it.

tmux new -s nim

I have a metal cup with the bottom missing and the top sealed. How can I use this cup?

Let’s continue where we stopped, look at our last session conversation and also read gwill-tech-theme-workflow skill, changelog and readme for guidance, our last session was incomplete so find out where u stopped and finish up