Conventions for writing knowledge-base files for an existing agent
Scanned 9/1/2026
Install to Claude Code
npx -y skills add phenobarbital/ai-parrot --skill kb-writer --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Kb Writer?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/phenobarbital-kb-writer)More formats (shields.io, HTML) on the badges page.
---
name: kb-writer
description: Conventions for writing knowledge-base files for an existing agent
triggers:
- write a kb file
- add knowledge
- update knowledge base
- kb file
category: agent-development
---
# KB Writer
Knowledge-base files give an existing agent extra reference material at
answer time. They live flat under `AGENTS_DIR/<agent_name>/kb/` — no
subdirectories.
## Rules
1. The target agent must already exist — check with
`list_existing_agents` first; if it doesn't, build the agent (see the
agent-builder skill) before writing KB content for it.
2. Filenames must be a single path segment (no `/`) with a `.md` or
`.txt` extension.
3. Write clear, self-contained sections — the retrieval layer chunks
these files, so avoid content that only makes sense with hidden
context from earlier in the document.
4. Call `write_kb_file(agent_name, filename, content)`. The response
always reports `reload_required: true` — tell the user the agent must
be reloaded (or restarted) before the new content is used; this skill
never triggers a reload itself.
Keep each file focused on one topic — prefer several small, well-titled
files over one large dump.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!
Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.
**Complete production-ready guide for Google Gemini embeddings API** This skill provides comprehensive coverage of the `gemini-embedding-001` model for generating text embeddings, including SDK usage, REST API patterns, batch processing, RAG integration with Cloudflare Vectorize, and advanced use cases like semantic search and document clustering. ---
Use when a repo needs CodeGraph plus ast-grep for Codex MCP setup, exploration, impact analysis, structural search, or safe refactor planning.
Interview, source-challenge, verify, save, and ADR-gate fuzzy coding requests into Codex-ready implementation specs. Use when a feature, bugfix, refactor, migration, repo-wide change, or architecture task needs user-verified requirements, source-backed decisions, durable architecture decisions, acceptance criteria, validation commands, rollout notes, saved spec/ADR files, and a Codex execution prompt. Do not use when already fully specified or when the user wants direct implementation now.