compound-docs

A documentation workflow for recording solved coding problems in searchable files so they can be reused later.

In plain words
What is it for?
It helps document solutions, search past solutions, promote repeated fixes into shared patterns, and validate their YAML metadata.
Why use it?
It prevents teams from solving the same problem repeatedly and helps them find earlier fixes by keyword, tag, or problem type.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/nth5693/gemini-kit/compound-docs
Any agent
npx skills add nth5693/gemini-kit --skill compound-docs
Clone the repo
git clone --depth 1 https://github.com/nth5693/gemini-kit

Made for: Claude Code, Codex.

Per session 10 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 744 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00010 $0.00744
Opus 5 $0.00005 $0.00372
Sonnet 5 $0.00002 $0.00149
Haiku 4.5 $0.00001 $0.00074

Measured 3d ago against content hash 1c3b6401c330, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

compound-docs scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

skills/compound-docs/SKILL.md · 134 lines

How it starts

The opening of the file, as written. The whole thing — 134 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Compound Documentation Skill

Manages solution documentation in docs/solutions/ for knowledge persistence across sessions.

Overview

When you solve a problem, document it so the solution can be found and reused later.

What do you want to do?

  1. Document a solution → Use /compound workflow
  2. Find existing solutions → See "Searching Solutions" below
  3. Promote to pattern → See "Pattern Promotion" below
  4. Validate schema → See "YAML Validation" below

Instrumentation

# Log usage when using this skill
./scripts/log-skill.sh "compound-docs" "manual" "$$"

Searching Solutions

# Search by keyword
grep -r "keyword" docs/solutions/

# Search by tag
grep -l "tags:.*performance" docs/solutions/**/*.md

# Search by problem type
ls docs/solutions/performance-issues/

Pattern Promotion

When an issue occurs 3+ times:

  1. Search for similar solutions:

    grep -r "similar terms" docs/solutions/
    
  2. Add to critical patterns:

    # Edit docs/solutions/patterns/critical-patterns.md
    
  3. Format:

    ### Pattern #{N}: {Name}
    
    **Problem:** {What goes wrong}
    
    **❌ WRONG:**
    ```code
    {incorrect}
    

    ✅ CORRECT:

    {correct}
    

YAML Validation

All solutions must have valid frontmatter:

---
date: "YYYY-MM-DD"
problem_type: "{from schema.yaml}"
severity: "critical|high|medium|low"
symptoms:
  - "symptom 1"
root_cause: "{from schema.yaml}"
tags:
  - tag1
---

See docs/solutions/schema.yaml for valid enum values.

Documenting Failures & Learnings

When documenting solutions, capture the full journey:

What to Include

  • ✅ Failed attempts with explanations
  • ✅ Incorrect assumptions and corrections
  • ✅ Key insights from mistakes
  • ✅ Course corrections and why
  • ✅ Time investment (helps prioritize future similar issues)

Example Pattern

## Investigation Steps

| Attempt | Hypothesis | Result |
|---------|------------|--------|
| 1. Check API logs | API might be timing out | ❌ No timeouts found |
| 2. Review database queries | N+1 query suspected | ❌ Queries were optimized |
| 3. Check cache invalidation | Cache might be stale | ✅ Found cache not clearing |

## Lessons Learned

### Mistakes Made
- **Assumption:** Assumed the API was the bottleneck
  - **Why wrong:** Didn't check cache layer first
  - **Correction:** Should always check cache before API

### Key Breakthrough
- **Insight:** Realized cache invalidation logic was only running on UPDATE, not DELETE
  - **Impact:** Led directly to the solution

Read the full file on GitHub · 134 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 134 lines · 10 tokens per session scan A 1c3b6401c330

Subscribe to this mod's changes

compound-docs is a skill published in the GitHub repository nth5693/gemini-kit (375 stars, last pushed 5mo ago), licensed MIT. It adds 10 tokens to every session and 744 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

triage-issue

Investigate a reported bug to root cause, then emit a TDD-shaped fix plan as an issue artifact. Trigger when the user reports a bug, says "triage", asks for issue investigation, or wants a fix plan before code changes.

OutlineDriven/odin-gemini-cli-extension · 54 tokens

minimalist-general

Subtraction-first thinking for non-coding tasks: writing, planning, research, summarizing, decision-making, or any request that isn't producing code. Same discipline as the minimalist coding skill — question whether the ask is even needed, reuse what already exists, do the smallest thing that fully answers it …

DivyeshJayswal/minimalist · 160 tokens

minimalist

Subtraction-first engineering for any coding task. Use when writing, fixing, refactoring, reviewing, or designing code; when choosing dependencies; or whenever the user asks for minimalist, less code, simplest thing, YAGNI, or complains about bloat. Prefer deletion, existing code, stdlib, and native platform features…

DivyeshJayswal/minimalist · 77 tokens

minimalist-audit

Audit a codebase or directory for deletion candidates: dead code, unused dependencies, single-use abstractions, config that never varies, and duplicated helpers. Use when the user says "minimalist audit" or asks what can be deleted from a project.

DivyeshJayswal/minimalist · 55 tokens

minimalist-gain

Report what minimalist actually measured in this session or project — LOC avoided, scope rejected, dependencies declined. Use when the user says "minimalist gain", "what did you save", or asks for the savings report.

DivyeshJayswal/minimalist · 48 tokens

minimalist-review

Review code, a diff, or a PR strictly for bloat: unrequested abstractions, dead scope, dependency creep, symptom-patching, and drive-by changes. Use when the user says "minimalist review", asks "is this over-engineered?", or wants a leanness review of a change.

DivyeshJayswal/minimalist · 67 tokens