Arm a prewalk run: explore and plan in the root session, land the first verified edit, then persist a Stop checkpoint for handoff.
Scanned 9/6/2026
Install to Claude Code
npx -y skills add TerenceLiu98/prewalk --skill prewalk --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Prewalk?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/terenceliu98-prewalk)More formats (shields.io, HTML) on the badges page.
---
name: prewalk
description: "Arm a prewalk run: explore and plan in the root session, land the first verified edit, then persist a Stop checkpoint for handoff."
---
# Prewalk
Arm the current session:
```bash
python3 "${CLAUDE_PLUGIN_ROOT}/hooks/_arm.py" arm "$CLAUDE_SESSION_ID" "$ARGUMENTS"
```
Options must precede task text: `--preset <name>` selects a preset and
`--fast` enables automatic handoff after validation (`--no-pause` is a legacy
alias). The active root session is the planner; Prewalk never changes it.
If the task is clearly one or two small edits, finish it directly. Otherwise:
1. Explore the relevant entry points, configuration, tests, and existing patterns.
2. Create at most the configured number of todos. Every real todo names concrete
files or behavior and includes a test/build/verify/check criterion.
3. Complete and verify only real task #1.
4. Update the todo snapshot so task #1 is `completed`; keep only real work in it.
5. Stop with this exact structured packet. Do not start task #2.
```text
Goal:
Files Read:
Constraints And Existing Patterns:
Full Todo List:
Task 1 Changes:
Verification Already Run:
Remaining Work:
Risks / Do Not Repeat:
```
Do not claim that handoff occurred. `/prewalk:pw-go` requests it; the bound
executor's `SubagentStop` hook confirms its final result.
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!
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.
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', ...
Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.
**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. ---
Recovers prior coding-agent session context by running `catchup <agent> --since-compact`, which extracts a clean summary of a previous Codex, Claude Code, Antigravity, OpenCode, or Pi Agent session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", or asks to recover/summarize a previous session before continuing. Do NOT use for the current conversation, git history, or any non-agent log.