A research-first workflow for starting projects in an unfamiliar area. It moves through observing, understanding, choosing an approach, and then creating the project.

Use it when starting a new app, library, command-line tool, skill, or other project involving a new domain or unfamiliar dependency. It helps gather sources, query local knowledge, and record findings before implementation.

It helps prevent coding before the needed background, local files, and outside sources are understood. The slower mode can require research before files may be changed.

oodc — Skill for Claude Code and Codex by mr-shaper

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.

Any agent

npx skills add mr-shaper/keel-harness --skill oodc

Clone the repo

git clone --depth 1 https://github.com/mr-shaper/keel-harness

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/mr-shaper/keel-harness/oodc/github.svg)](https://agentmods.dev/skills/mr-shaper/keel-harness/oodc)

Your own site

<a href="https://agentmods.dev/skills/mr-shaper/keel-harness/oodc"><img src="https://agentmods.dev/badge/skills/mr-shaper/keel-harness/oodc/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/mr-shaper/keel-harness/oodc"><img src="https://agentmods.dev/badge/skills/mr-shaper/keel-harness/oodc.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 110 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,361 The whole file, excluding the scripts and references it only reads on demand.

Security scan B 1 finding. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00110	$0.01361
Opus 5	$0.00055	$0.00681
Sonnet 5	$0.00022	$0.00272
Haiku 4.5	$0.00011	$0.00136

Measured 11d ago against content hash 1ae7a298147a, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade B, and why

oodc scanned grade B with 1 finding 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 11d 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.

Enumerates other installed skillsmediumAgent snooping

Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.

3. `Bash: ls ~/.claude/skills/shelf/ | head -20`

plugins/oodc/skills/oodc/SKILL.md · 95 lines

How it starts

The opening of the file, as written. The whole thing — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.

OODC — Observe → Orient → Decide → Create

KB: python3 ~/.claude/plugins/tacit-kb/scripts/kb.py Slow-loop: .oodc-state-{project} blocks Write/Edit until CREATE. Read ~/.claude/workflows/oodc-loop.md.

Mode

Ask user: "Fast or slow?"

Slow: new domain / upstream / "deep research". Run: Bash: echo -e "step=OBSERVE\nproject={name}\nstarted=$(date -u +%FT%TZ)" > ~/.claude/.oodc-state-{name}
Fast: familiar domain. No state file.

OBSERVE

Tool routing: web-access=GitHub / bird=X / notebooklm=NLM. Details: references/observe-protocol.md.

Slow: 3 parallel Agents. NLM: notebook → ≥30 sources → 6 Q&A.

Fast: Run these commands:

Bash: python3 ~/.claude/plugins/tacit-kb/scripts/kb.py wakeup --tokens 400
Bash: python3 ~/.claude/plugins/tacit-kb/scripts/kb.py query "{project_keywords}" --domain {domain} --top 3 (requires tacit-kb plugin — optional)
Bash: ls ~/.claude/skills/shelf/ | head -20
Read related local doc directories

NLM answers only. Down → [⚠️ non-NLM].

Required Output [OBSERVE_RESULT]: github: {findings} | x: {insights} | nlm: {Q&A} | local: {assets} | source_count: N

Done = source_count ≥ 3 (fast) / ≥ 20 (slow). Each field: concrete finding, not "no results". Empty field (slow) → do not proceed. Slow: Bash: perl -pi -e 's/step=.*/step=ORIENT/' ~/.claude/.oodc-state-{project}

ORIENT

See references/orient-playbook.md.

Assumptions: 3 × (text / evidence / counter / confidence)
Experts: Use Skill tool — NOT text mention, TOOL CALL. Run: Skill tool, skill name = "{perspective-name}" (e.g. feynman-perspective). Cap: 4. Fallback if Skill fails: Bash: head -40 ~/.claude/skills/{name}/SKILL.md + [⚠️ fallback]
PUA: Run: Skill tool, skill name = "pua:p10" — apply three questions.

Read the full file on GitHub · 95 lines

Files

What ships with it

4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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.

11d ago First seen · 95 lines · 110 tokens per session scan B 1ae7a298147a

Subscribe to this mod's changes

oodc is a skill published in the GitHub repository mr-shaper/keel-harness (5 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 110 tokens to every session and 1,361 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it B with 1 finding (enumerates other installed skills). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Claude Code Planning Research agentic-engineering ai-agents developer-tools karpathy shell

mcpx

Use this skill when interacting with MCP servers via CLI. Prefer mcpx over direct MCP SDK/protocol calls for tool discovery, schema inspection, invocation, and Unix-style output composition.

lydakis/mcpx · 39 tokens

cspell-triage

Use when cspell reports unknown words or forbidden words in a repository, when adding entries to a cspell dictionary, or when a cspell config needs setting up, consolidating, or auditing.

AndrewDongminYoo/cc-agents-kit · 45 tokens

fix-osv-vulnerabilities

Use when osv-scanner or trunk check reports dependency vulnerabilities (GHSA-) in a pnpm/npm/yarn project and you need to triage and fix them.

AndrewDongminYoo/cc-agents-kit · 41 tokens

config-gc

Garbage collection for the active Claude config root — scans skills, memory, hooks, permissions, MCP servers, and caches for redundant, stale, or orphaned items, then walks the user through confirm-each-deletion cleanup. Use on "clean up my config", "config GC", "too many skills", "my .claude is bloated", or a…

AndrewDongminYoo/cc-agents-kit · 81 tokens

wayfinder

Chart an effort too big for one session as a map of decision tickets under docs/plans/, then resolve one ticket per session until the way to the destination is clear.

AndrewDongminYoo/cc-agents-kit · 37 tokens

ci-babysit

Use after pushing a branch or opening a PR when CI needs monitoring or authorized make-green work. Keeps watch-only requests read-only, requires explicit current-request or approved-plan authority before reruns or local repairs, and stops honestly when a human or logic change is needed. Works in any ecosystem.

AndrewDongminYoo/cc-agents-kit · 64 tokens

oodc