| name | sandbox-explain |
| description | Create or update explanatory Obsidian notes in Flo's agent sandbox. Use when the user asks to create a sandbox file, agent handbook note, explanation note, visual explainer, or communication artifact in the agent sandbox; when they mention the Agent Sandbox, sandbox folder, INDEX.md, or helping agents understand a project/concept. |
Sandbox Explain
Create clear explanatory notes in Flo's Obsidian agent sandbox.
Fixed Target
- Use Obsidian vault
TheVault.
- Use sandbox root
_AgentSandbox_/.
- Do not use other vaults for this workflow.
- Create shared personal skill outputs under the vault, not inside a project repo.
Required Companion Skills
- Use
obsidian-cli for vault lookup, reads, and writes when Obsidian interaction is needed.
- Use
obsidian-markdown for Obsidian-valid Markdown, wikilinks, callouts, embeds, and Mermaid.
- Ignore
obsidian-markdown's generic frontmatter guidance for this workflow.
- Use
apply-style for Flo's documentation style.
- Do not duplicate Flo's full style rules in this skill; defer to
apply-style and the base agent instructions.
Hard Formatting Rules
- Do not write YAML frontmatter in sandbox notes or sandbox
INDEX.md files.
- No opening
--- property block.
- No
title, tags, created, status, or similar metadata properties.
- Put useful context in normal Markdown only when it helps future reading.
- Do not repeat the filename as the first
# H1.
- The note title lives in the filename.
- Start with useful content instead:
- a callout
- a
## section
- a Mermaid diagram
- a short result / mental-model statement
- Use an H1 only when the user explicitly asks for one or when it is not just the filename repeated.
Workflow
- Identify the topic or project the note should explain.
- Inspect
_AgentSandbox_/ first-level folders.
- Choose the folder:
- Reuse an existing emoji-prefixed folder when it clearly matches the topic.
- Create a new emoji-prefixed folder only when no good candidate exists.
- Use a semantic emoji that helps scanning.
- Ensure the folder has
INDEX.md.
- Create it if missing.
- Update it if the new note should be listed.
- Keep it short: purpose, key notes, and current status.
- Create one focused explanation note unless the user asks for more.
- Apply Flo's style through
apply-style.
- Verify Obsidian can read the created or updated note.
Default Output Shape
Use this default unless the user asks otherwise:
- Folder:
_AgentSandbox_/{emoji} {Topic}/
- Index:
_AgentSandbox_/{emoji} {Topic}/INDEX.md
- Topic note:
_AgentSandbox_/{emoji} {Topic}/{short-title}.md
INDEX.md Contents
Keep INDEX.md as a lightweight entrypoint:
- What this folder is for
- Important notes in the folder
- Current status or next useful read
- Links using Obsidian wikilinks
Do not turn INDEX.md into the full explanation unless the user explicitly wants a single-file folder.
Explanation Note Guidance
Default to a note that helps another agent or future Flo understand the work quickly:
- Start with the result or mental model, not a repeated title.
- Use
## headings for sections by default.
- Include a Mermaid diagram when it materially improves understanding.
- Use concrete examples from the current project or artifact.
- Clearly separate:
- what exists now
- how it works
- what is not built yet
- what to check next
- Use code fences only for real code, commands, config, or data shapes.
Folder Reuse Heuristic
Prefer reuse when:
- Folder name contains the project/topic name.
- Folder contents already cover the same system or investigation.
- The user says "this sandbox" or "the existing folder".
Create a new folder when:
- Existing folders are unrelated.
- Reuse would mix unrelated projects.
- The user asks for a new sandbox area.
Safety
- Do not delete or rename existing sandbox notes unless explicitly asked.
- Do not move notes between folders without asking.
- Do not create Codex-local skill adapters under
~/.codex/skills.
- If creating or updating this skill itself, use
/Users/flo/.agents/skills.