skill CX Codex

reset — Claude Code skill: what it does, how to install

Q: How do I install reset in Claude Code?

Run `npx skills add NeoLabHQ/context-engineering-kit --skill reset -a claude-code`. The tool places the skill definition under `~/.claude/skills/reset/SKILL.md`.

Q: What is the difference between soft and hard reset?

Soft reset logs session metadata in `.fpf/sessions/` and archives files instead of destroying them. Hard reset permanently deletes markdown files across `.fpf/knowledge/`, `.fpf/evidence/`, and `.fpf/decisions/`.

Clears or archives the active FPF reasoning cycle so you can start a fresh analytical session.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓

npx skills add NeoLabHQ/context-engineering-kit --skill reset -a claude-code

Open on GitHub

What reset does

This skill manages the FPF reasoning cycle stored in your .fpf directory. It gives your agent a structured procedure to restart work without losing track of past ideas or cluttering the workspace.

Three reset modes are available. The soft reset writes a summary to .fpf/sessions/ and moves active hypotheses into an archive folder. The hard reset clears out all markdown files from knowledge levels, evidence, and decisions. Because hard reset wipes those files permanently, the agent executes it only upon explicit confirmation.

The third mode resets decisions while keeping gathered knowledge intact. It downgrades validated L2 hypotheses back to L1 for a re-audit, or moves L1 hypotheses back to L0 for verification.

What it helps with

archive active session hypotheses into .fpf/sessions/
wipe knowledge and evidence directories for a clean restart
move L2 hypotheses back to L1 for re-auditing
re-verify initial assumptions by rolling L1 items back to L0

Example prompts for the agent

Perform a soft reset on the current FPF cycle and archive all hypotheses
Do a hard reset and clear all files in the .fpf directories
Reset decisions by moving L2 hypotheses back to L1 for re-evaluation

How to install reset

1Skills CLI (Vercel)

npx skills add NeoLabHQ/context-engineering-kit --skill reset -a claude-code

2Manually: put the files in place

~/.claude/skills/reset/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/reset/

1Skills CLI (Vercel)

npx skills add NeoLabHQ/context-engineering-kit --skill reset -a codex

2Manually: put the files in place

.agents/skills/reset/SKILL.md

For all projects use ~/.agents/skills/

1Skills CLI (Vercel)

npx skills add NeoLabHQ/context-engineering-kit --skill reset -a cursor

2Manually: put the files in place

.cursor/skills/reset/SKILL.md

Cursor also reads .agents/skills/ and .claude/skills/

1GitHub CLI

gh skill install NeoLabHQ/context-engineering-kit reset

2Manually: put the files in place

.github/skills/reset/SKILL.md

1Skills CLI (Vercel)

npx skills add NeoLabHQ/context-engineering-kit --skill reset -a gemini-cli

2Platform CLI

gemini skills install https://github.com/NeoLabHQ/context-engineering-kit --path skills/reset

Manually: put the files in place

.devin/skills/reset/SKILL.md

The legacy .windsurf/skills/ path still works

Skills CLI (Vercel)

npx skills add NeoLabHQ/context-engineering-kit --skill reset -a opencode

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about reset

How do I install reset in Claude Code?

Run npx skills add NeoLabHQ/context-engineering-kit --skill reset -a claude-code. The tool places the skill definition under ~/.claude/skills/reset/SKILL.md.

What is the difference between soft and hard reset?

Soft reset logs session metadata in .fpf/sessions/ and archives files instead of destroying them. Hard reset permanently deletes markdown files across .fpf/knowledge/, .fpf/evidence/, and .fpf/decisions/.

What does the decision reset option do?

It keeps existing notes but steps back hypothesis stages. It moves items from L2 down to L1 or from L1 down to L0 so the agent can re-evaluate them.

ai-shaped-readiness-advisor deanpeters A guided assessment of how your product team really uses AI: speeding up existing tasks or rebuilding the work around it, plus what capability to build next. CX Codex +1 2,700installs context-engineering addyosmani Helps you set up agent context: which rules, specs and files to feed the agent, and when. Use it at the start of a session, when output quality drops, or when switching tasks. CX Codex +1 61,600installs consciousness-council K-Dense-AI Runs a council of thinking archetypes over a hard choice, keeps facts apart from assumptions and values, and returns a conditional recommendation. CX Codex +1 1,900installs sequential-thinking mrgoonie Breaks complex problems into iterative reasoning steps with dynamic scope and hypothesis branching. Built for architecture planning and deep debugging. CX Codex +1 1,800installs analysis microchipgnu Turns a contested situation into a structured brief: the hidden bet, the real disagreement, who pays. Built for news, conflicts and decisions that need a judgment, not a summary. CX Codex +1 1,400installs thinking-first-principles tjboudreaux A skill that makes the agent separate real constraints from convention and rebuild the simplest solution. Useful when a task looks too costly or impossible. CX Codex +1 708installs

From the same set55

why Conducts iterative Five Whys root cause analysis for technical bugs and incidents. Helps uncover process and system gaps instead of settling for superficial patches. CX Codex +1 1,200installs review-local-changes Reviews uncommitted working-tree changes and returns a list of issues with improvement suggestions. Handy right before you commit. CX Codex +1 1,200installs git-notes Attaches notes to commits without touching their SHA. Useful for marking reviews, test results or adding context after the fact. CX Codex +1 1,100installs setup-context7-mcp This skill walks the agent through setting up the Context7 MCP server and writes the library IDs into CLAUDE.md so code follows current docs. CX Codex +1 1,100installs status Reports what is in the FPF knowledge base right now: hypothesis counts per layer, evidence freshness, and recent decisions. CX Codex +1 1,100installs query Searches hypotheses and decisions under .fpf/ and shows each one with its layer, dependencies and R_eff score. CX Codex +1 1,100installs create-workflow-command This skill generates multi-step workflow commands that delegate execution to sub-agents using dedicated task files, preventing context bloat in complex runs. CX Codex +1 1,100installs decay Checks expiry dates on evidence files in .fpf/evidence/ and offers refresh, deprecate or waive actions for decisions built on stale data. CX Codex +1 1,200installs