Skills for Real Engineers. Straight from my .agents directory.
think
Agent skill fromAgriciDaniel/claude-obsidian
Apply the Fable-derived 10-stage OBSERVE, OBSERVE, LISTEN, THINK, CONNECT, CONNECT, FEEL, ACCEPT, CREATE, GROW loop to consequential or ambiguous reasoning and decisions. Use for think this through, deep think, architecture review, postmortem, tradeoff analysis, or challenges to assumptions. This is a reasoning review, not a deterministic vault health check or repair workflow.
How to install think
npx skills add AgriciDaniel/claude-obsidian# or add only this skill to Claude Code: copy the folder
skills/think/
# into
~/.claude/skills/think/ (all projects)
.claude/skills/think/ (one project)
1# Think23Use the ten stages as a loop of attention, not a ceremonial transcript. Scale4the depth to novelty, stakes, uncertainty, and irreversibility. Keep the final5reasoning summary concise and evidence-bearing.67This skill is read-only. It may inspect available evidence, but it does not8edit files, execute a vault mutation, commit Git, or publish anything. If the9CREATE stage recommends a change, compose with the appropriate mutation skill10after the user authorizes that operation.1112## Establish an evidence frame1314Before the loop, separate:1516- **Observed:** directly supported by an input, source, test, or measurement.17- **Inferred:** a conclusion drawn from observations; state the bridge.18- **Unknown:** missing evidence that could change the decision.19- **Preference:** a user or designer choice, not a factual claim.2021Attach citations or concrete locators when sources exist. Never fabricate a22quotation, locator, test result, consensus, or confidence level.2324## The ten stages2526### 1. OBSERVE: external2728Read the actual inputs, current state, constraints, failures, and prior work.29Distinguish what was inspected in full from what was sampled or assumed.3031### 2. OBSERVE: internal3233Name the biases most likely to distort this decision: anchoring, ownership,34novelty, familiarity, urgency, sunk cost, confirmation, or a desired verdict.35Recalibrate confidence to the evidence available.3637### 3. LISTEN3839Recover the user's intent, success condition, constraints, and unstated stakes.40…
Excerpt shown under the repository's MIT license.
More skills in claude-obsidian
Similar skill repositories
An agentic skills framework & software development methodology that works.
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Public repository for Agent Skills
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
