elliotjlt/claude-skill-potions/breadcrumbs
Claude Code and agent skill by elliotjlt/claude-skill-potions
- community
Checking that the SKILL.md file is still in the repository.
Leaves notes for future Claude sessions in .claude/breadcrumbs.md. At session end or after significant work, records what was tried, what worked, what failed, and context for next time. On session start, reads breadcrumbs to understand history. Session-to-session memory without manual updates.
Install
npx vskill@latest install elliotjlt/claude-skill-potions/breadcrumbsSKILL.md
Breadcrumbs
<purpose> Claude sessions are stateless. You close the terminal, context is gone. Next session starts fresh with no memory of what was tried, what failed, or why certain approaches were abandoned. Breadcrumbs fixes this - notes you leave for future-you (or future-Claude) to find. </purpose>
When To Activate
<triggers> Read breadcrumbs (session start):
- Beginning of a new session
- User says "what were we working on?"
- User references previous session work
- Before starting work on a familiar codebase
Write breadcrumbs (during/end of session):
- After completing significant work
- After hitting a dead end or failed approach
- Before ending a session
- User says "remember this" or "note for next time"
- After discovering something important about the codebase
</triggers>
Instructions
File Location
Create .claude/ directory if it doesn't exist. Add to .gitignore if user prefers.
Reading Breadcrumbs (Session Start)
<read_breadcrumbs> At session start, check for breadcrumbs:
If found, summarize key points:
Don't read the whole file aloud - summarize what's relevant. </read_breadcrumbs>
Writing Breadcrumbs
<write_breadcrumbs> Append to breadcrumbs file (don't overwrite):
</write_breadcrumbs>
Breadcrumb Types
Discovery breadcrumb - Found something important:
Dead end breadcrumb - Tried something that failed:
Progress breadcrumb - Work in progress:
Context breadcrumb - Important background:
Breadcrumb Hygiene
Pruning old breadcrumbs:
- Keep last 10-15 entries
- Archive old ones to
.claude/breadcrumbs-archive.md - Delete truly obsolete info (completed features, resolved issues)
When to prune:
- File exceeds ~200 lines
- User asks to clean up
- Old breadcrumbs reference deleted code
Output Format
When reading:
When writing:
NEVER
- Overwrite the entire breadcrumbs file (append only)
- Include sensitive data (passwords, keys, secrets)
- Write breadcrumbs for trivial work
- Ignore existing breadcrumbs at session start
- Write vague notes ("worked on stuff")
ALWAYS
- Check for breadcrumbs at session start
- Record dead ends (future you will thank you)
- Include file paths for important discoveries
- Note the "why" not just the "what"
- Keep entries scannable (bullets, not paragraphs)
Example Session
Session 1 ends:
Session 2 starts:
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/breadcrumbs/SKILL.md
Popularity
- Repository stars
- 51
- Repository forks
- 2
- 7-day trend score
- 20
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 19 Mar 2026, 7 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/elliotjlt/claude-skill-potions/breadcrumbs)