memory-manage
Skill38 starsChanged 9 months agoMITView on GitHub
Manage agent memory stored as files. Record, recall, search, update, and analyze memory content. Supports exploratory analysis for building context before complex tasks.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- zhao-wuyan/ccw-command-explorertheir own version, first seen May 8, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- zhao-wuyan/ccw-command-explorer★ 45same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
Show its rules and text
We split this skill into the rules it gives an agent
20 rules. About 0 are common, 15 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈15
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Default memory file: memory.txt in the current working directory
- Very specific: line 14
If the memory file does not exist yet, create it automatically on the first record
- Distinctive: line 24
Read the full contents of the memory file using the Read tool and display it to the user.
- Distinctive: line 26
If the file does not exist, inform the user that no memory has been recorded yet.
- Distinctive: line 27
If the file is large (over 200 lines), provide a summary of its structure first, then ask if the user wants the full content.
- Distinctive: line 41
Use the current date and time for the timestamp
- Very specific: line 43
If the file does not exist, create it with a header line: # Memory File
- Distinctive: line 44
Use the Read tool to get existing content, then Write to append the new entry
- Distinctive: line 45
Do NOT overwrite existing content — always append
- Distinctive: line 53
Use the Read tool to find the content
- Distinctive: line 54
Use the Edit tool to perform the replacement
- Distinctive: line 61
Search the memory file for matching content.
- Distinctive: line 63
Use the Grep tool to find lines matching the query
- Distinctive: line 65
If no matches are found, suggest alternative search terms based on the query
- Distinctive: line 71
Remove specific content from the memory file.
- Distinctive: line 73
Use the Read tool to locate the content
- Distinctive: line 74
Use the Edit tool to remove the matching section (including its timestamp header if it is a complete entry)
- Distinctive: line 84
Read the entire memory file using the Read tool
- Very specific: line 116
If /memory is invoked without arguments or with an unrecognized command:
- Very specific: line 118
Default to analyze if the memory file exists, or inform the user to start with record if it does not
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Memory Management Skill
A Claude Code skill that implements the Memory as File concept — treating conversation knowledge as persistent files, managed entirely through the Coding Agent's native tools.
Default Behavior
- Default memory file:
memory.txtin the current working directory - The user can specify a custom file path as the last argument to any command
- If the memory file does not exist yet, create it automatically on the first
record
Commands
Parse $ARGUMENTS to determine which operation to perform. The first word is the command, the rest are arguments.
recall — Read Memory
Usage: /memory recall [file]
Read the full contents of the memory file using the Read tool and display it to the user.
- If the file does not exist, inform the user that no memory has been recorded yet.
- If the file is large (over 200 lines), provide a summary of its structure first, then ask if the user wants the full content.
record — Append to Memory
Usage: /memory record <content> [file]
Append new information to the memory file.
Format each entry as:
## [YYYY-MM-DD HH:MM] <short topic>
<content>- Use the current date and time for the timestamp
- Derive a short topic label (3-5 words) from the content
- If the file does not exist, create it with a header line:
# Memory File - Use the Read tool to get existing content, then Write to append the new entry
- Do NOT overwrite existing content — always append
update — Modify Memory
Usage: /memory update <old> -> <new> [file]
Replace specific content in the memory file.
- Use the Read tool to find the content
- Use the Edit tool to perform the replacement
- Confirm the change to the user by showing the before and after
search — Search Memory
Usage: /memory search <query> [file]
Search the memory file for matching content.
- Use the Grep tool to find lines matching the query
- Display matching entries with surrounding context (2 lines before and after)
- If no matches are found, suggest alternative search terms based on the query
forget — Remove from Memory
Usage: /memory forget <content> [file]
Remove specific content from the memory file.
- Use the Read tool to locate the content
- Use the Edit tool to remove the matching section (including its timestamp header if it is a complete entry)
- Confirm what was removed
analyze — Exploratory Analysis
Usage: /memory analyze [file]
This is the most important command. Perform a comprehensive exploratory analysis of the memory file to build context for subsequent work.
Steps:
- Read the entire memory file using the Read tool
- Produce a structured analysis report with the following sections:
## Memory Analysis Report
### Summary
One-paragraph overview of what the memory contains — its scope, depth, and overall theme.
### Topics
Bulleted list of distinct topics and themes found in the memory, ordered by frequency or importance.
### Key Entities
People, projects, tools, technologies, and decisions mentioned. Group by category.
### Timeline
Chronological progression of events or knowledge, if timestamps or temporal references exist.
### Relationships
Connections between topics — how different pieces of memory relate to each other.
### Knowledge Gaps
What information is missing, incomplete, or unclear. What questions remain unanswered.
### Suggested Next Steps
Concrete, actionable recommendations for what the agent (or user) could do next based on the memory content. These should be specific enough to serve as task descriptions.- This report is designed to give the agent (or a sub-agent) a basic info foundation — enough context to understand the situation and continue working on downstream tasks without needing to re-read the full memory file.
When No Command Is Specified
If /memory is invoked without arguments or with an unrecognized command:
- Display a brief help message listing all available commands with one-line descriptions
- Default to
analyzeif the memory file exists, or inform the user to start withrecordif it does not
Design Principles
- Files are memory — all knowledge is persisted as human-readable text files
- Native tools only — use Read, Write, Edit, Grep, and Glob; no external dependencies
- Human-auditable — users can open and edit memory files directly at any time
- Timestamp everything — every recorded entry gets a timestamp for chronological tracking
- Analyze before acting — the
analyzecommand provides the foundation for informed decision-making
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

