How do I install the Tmp Snapshot skill?

Run npx skillmds@latest add otrebu/tmp-snapshot in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Tmp Snapshot safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Tmp Snapshot?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Tmp Snapshot free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Tmp Snapshot?

otrebu (@otrebu) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Tmp Snapshot by otrebu — Agent Skill

Name: Tmp Snapshot
Author: otrebu

Tmp Snapshot

Save content from the current conversation to /tmp/<descriptive-kebab-name>.md, then return the absolute path and an outline (every header with a one-line summary). Use when the user says "save this to /tmp", "snapshot this", "dump the X part", "stash this", "save the discussion about Y", or otherwise asks to drop content from the chat into a quick scratch markdown file.

otrebu 3b765a3 2 files · 2.5 KB Updated Sep 17, 2026

File contents

tmp-snapshot

Save content the user asked for into /tmp/<name>.md, then reply with the path and an outline. Nothing more.

Steps

Pick the content. Use what the user pointed at:
- A topic from this conversation ("save the auth part") → extract just that slice.
- The whole chat ("save everything we just discussed") → summarize the substantive bits.
- Text they pasted into their message → use it verbatim.
If the slice is ambiguous, ask once before writing.
Pick a kebab-case filename. 2–4 words, descriptive of the content, no dates unless they matter to the topic. Path is always /tmp/<name>.md — never a subfolder unless the user specifies one. Overwrite if it exists; /tmp is scratch.
Write the file as standalone markdown: one # H1 title, ## for sections, ### for sub-sections. It should read cleanly without the chat as context. No YAML frontmatter unless the user asks.
Reply with the path + outline. No preamble, no closing summary. Just those two things.

Outline format

A tree of every header in the saved file, each followed by one short sentence summarizing what's in that section. Indent by header depth.

Example reply

/tmp/auth-token-rotation.md

# Auth token rotation — Migrate from server sessions to JWT with refresh tokens.
  ## Current state — Sessions in Postgres, 24h expiry.
  ## Target — Short-lived JWT access + HttpOnly refresh cookie.
  ## Migration — Three-phase rollout plan.
    ### Phase 1 — Dual-write to both stores.
    ### Phase 2 — Read new, fall back to old.
    ### Phase 3 — Drop legacy session table.

Notes

Keep each summary to one line. The outline should be glanceable, not a re-paste of the doc.
If the user asks for a different directory (e.g. /tmp/notes/), honor it — but default is bare /tmp/.
Don't ask "are you sure?" before overwriting — that's what /tmp is for.

---name: tmp-snapshotdescription: Save content from the current conversation to /tmp/<descriptive-kebab-name>.md, then return the absolute path and an outline (every header with a one-line summary). Use when the user says "save this to /tmp", "snapshot this", "dump the X part", "stash this", "save the discussion about Y", or otherwise asks to drop content from the chat into a quick scratch markdown file.---# tmp-snapshotSave content the user asked for into `/tmp/<name>.md`, then reply with the path and an outline. Nothing more.## Steps1. **Pick the content.** Use what the user pointed at: - A topic from this conversation ("save the auth part") → extract just that slice. - The whole chat ("save everything we just discussed") → summarize the substantive bits. - Text they pasted into their message → use it verbatim. If the slice is ambiguous, ask **once** before writing.2. **Pick a kebab-case filename.** 2–4 words, descriptive of the content, no dates unless they matter to the topic. Path is always `/tmp/<name>.md` — never a subfolder unless the user specifies one. Overwrite if it exists; `/tmp` is scratch.3. **Write the file** as standalone markdown: one `#` H1 title, `##` for sections, `###` for sub-sections. It should read cleanly without the chat as context. No YAML frontmatter unless the user asks.4. **Reply with the path + outline.** No preamble, no closing summary. Just those two things.## Outline formatA tree of every header in the saved file, each followed by **one short sentence** summarizing what's in that section. Indent by header depth.### Example reply`/tmp/auth-token-rotation.md````# Auth token rotation — Migrate from server sessions to JWT with refresh tokens. ## Current state — Sessions in Postgres, 24h expiry. ## Target — Short-lived JWT access + HttpOnly refresh cookie. ## Migration — Three-phase rollout plan. ### Phase 1 — Dual-write to both stores. ### Phase 2 — Read new, fall back to old. ### Phase 3 — Drop legacy session table.```## Notes- Keep each summary to one line. The outline should be glanceable, not a re-paste of the doc.- If the user asks for a different directory (e.g. `/tmp/notes/`), honor it — but default is bare `/tmp/`.- Don't ask "are you sure?" before overwriting — that's what `/tmp` is for.

Frequently asked questions