claude-memory-bank-cli 1.1.0 → 1.2.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "claude-memory-bank-cli",
3
- "version": "1.1.0",
3
+ "version": "1.2.0",
4
4
  "description": "Fully automatic persistent memory system for Claude Code. One command, zero maintenance.",
5
5
  "main": "src/init.js",
6
6
  "bin": {
@@ -1,59 +1,61 @@
1
- # Project Memory Bank
2
-
3
- This project uses an automatic persistent memory system. Claude MUST follow the rules below every single session.
4
-
5
- ## Memory System Rules
6
-
7
- ### First Run (Empty Templates)
8
- If memory bank files still contain placeholder text (like `[Project name will be auto-filled by Claude]`), this is a first run. You MUST:
9
- 1. Scan the entire project — read package.json, source files, folder structure, configs, README, etc.
10
- 2. Auto-populate ALL memory bank files with real information extracted from the codebase
11
- 3. Fill in `projectbrief.md` with the project's actual name, goals, and scope
12
- 4. Fill in `techContext.md` with the real tech stack, dependencies, and commands
13
- 5. Fill in `productContext.md` with what the project does and why
14
- 6. Fill in `patterns.md` with code conventions you detect in the existing code
15
- 7. Leave `activeContext.md`, `progress.md`, `decisions.md`, `troubleshooting.md` ready for use
16
- 8. Do this BEFORE responding to the user's first message
17
-
18
- ### On Every Conversation Start
19
- 1. Read ALL files in `memory-bank/` before doing anything else
20
- 2. Check `memory-bank/activeContext.md` to understand what was last being worked on
21
- 3. Check `memory-bank/progress.md` to know current project status
22
- 4. If `memory-bank/extensions/` contains files, read those too
23
-
24
- ### On Every Task Completion
25
- 1. Update `memory-bank/activeContext.md` with what was just done
26
- 2. Update `memory-bank/progress.md` with current status
27
- 3. If a decision was made, add it to `memory-bank/decisions.md`
28
- 4. If a bug was fixed or discovered, update `memory-bank/troubleshooting.md`
29
- 5. If a new pattern or convention was established, update `memory-bank/patterns.md`
30
- 6. If the user expressed a preference, update `memory-bank/preferences.md`
31
- 7. If an external resource was referenced, update `memory-bank/references.md`
32
- 8. If tech stack or dependencies changed, update `memory-bank/techContext.md`
33
- 9. If project scope or goals changed, update `memory-bank/projectbrief.md` or `memory-bank/productContext.md`
34
-
35
- ### Overflow Rule
36
- - If any memory file exceeds ~200 lines, split overflow into `memory-bank/extensions/{filename}-{topic}.md`
37
- - Keep the main file as a summary with pointers to extension files
38
- - Extension files follow the same format as their parent file
39
-
40
- ### Auto-Cleanup
41
- - Remove outdated entries from `activeContext.md` when tasks are completed
42
- - Archive completed items in `progress.md` periodically
43
- - Remove stale references from `references.md` if links are confirmed dead
44
-
45
- ## Memory Bank Files
46
-
47
- | File | Purpose |
48
- |---|---|
49
- | `memory-bank/projectbrief.md` | Project scope, requirements, goals |
50
- | `memory-bank/productContext.md` | Why the project exists, user problems, UX goals |
51
- | `memory-bank/techContext.md` | Tech stack, dependencies, dev environment setup |
52
- | `memory-bank/activeContext.md` | Current focus — what's being worked on right now |
53
- | `memory-bank/progress.md` | Task tracking — done, in-progress, pending, blocked |
54
- | `memory-bank/decisions.md` | Every decision + reasoning + alternatives rejected |
55
- | `memory-bank/patterns.md` | Architecture patterns, code conventions, best practices |
56
- | `memory-bank/troubleshooting.md` | Bugs hit, fixes applied, lessons learned |
57
- | `memory-bank/preferences.md` | User preferences — communication style, design choices |
58
- | `memory-bank/references.md` | External links, APIs, tools, documentation |
59
- | `memory-bank/extensions/` | Overflow files when any main file exceeds ~200 lines |
1
+ # MANDATORY: Memory Bank System
2
+
3
+ > **IMPORTANT:** This project uses a memory bank. You MUST read and update it. This is NOT optional.
4
+
5
+ ## RULE 1: Read Memory Bank at Start — EVERY conversation
6
+
7
+ Before you respond to the user, you MUST read ALL of these files:
8
+ - `memory-bank/projectbrief.md`
9
+ - `memory-bank/productContext.md`
10
+ - `memory-bank/techContext.md`
11
+ - `memory-bank/activeContext.md`
12
+ - `memory-bank/progress.md`
13
+ - `memory-bank/decisions.md`
14
+ - `memory-bank/patterns.md`
15
+ - `memory-bank/troubleshooting.md`
16
+ - `memory-bank/preferences.md`
17
+ - `memory-bank/references.md`
18
+ - Any files in `memory-bank/extensions/`
19
+
20
+ Do NOT skip this step. Do NOT summarize from memory. READ the files.
21
+
22
+ ## RULE 2: First Run — If Files Are Empty, Populate Them
23
+
24
+ If any memory bank file contains placeholder text like `[Project name will be auto-filled by Claude]`, you MUST:
25
+ 1. Scan the entire project every source file, config, package.json, README, folder structure
26
+ 2. Write real project information into EVERY memory bank file
27
+ 3. Do this BEFORE responding to the user
28
+
29
+ The memory bank must NEVER remain empty after a conversation.
30
+
31
+ ## RULE 3: Update Memory Bank After EVERY Task
32
+
33
+ After completing any task (writing code, fixing a bug, answering a question, making a decision), you MUST immediately update the relevant memory bank files:
34
+
35
+ **Always update:**
36
+ - `memory-bank/activeContext.md` what you just did, what's next
37
+ - `memory-bank/progress.md` mark tasks done/in-progress/pending
38
+
39
+ **Update when relevant:**
40
+ - `memory-bank/decisions.md` — any choice made between alternatives
41
+ - `memory-bank/patterns.md` any new code pattern or convention
42
+ - `memory-bank/troubleshooting.md` — any bug found, fixed, or workaround used
43
+ - `memory-bank/preferences.md` any user correction, approval, or stated preference
44
+ - `memory-bank/references.md` — any external link, API, or tool mentioned
45
+ - `memory-bank/techContext.md` any dependency, tool, or setup change
46
+ - `memory-bank/projectbrief.md` — any scope or goal change
47
+ - `memory-bank/productContext.md` any product or UX goal change
48
+
49
+ **Do NOT wait until the end of the conversation. Update IMMEDIATELY after each task.**
50
+
51
+ ## RULE 4: Overflow Management
52
+
53
+ If any memory file exceeds ~200 lines:
54
+ 1. Create `memory-bank/extensions/{filename}-{topic}.md` with the overflow content
55
+ 2. Keep the main file as a summary with a pointer: `→ See extensions/{filename}-{topic}.md`
56
+
57
+ ## RULE 5: Auto-Cleanup
58
+
59
+ - Remove completed tasks from `activeContext.md`
60
+ - Remove stale or outdated information
61
+ - Keep files scannable — bullet points, not paragraphs
@@ -1,41 +1,41 @@
1
- # Memory Bank Auto-Update Rule
1
+ # MANDATORY: Auto-Update Memory Bank
2
2
 
3
- You MUST follow these rules automatically. No user command is needed.
3
+ > **You MUST follow these rules automatically. No user command is needed. No exceptions.**
4
4
 
5
- ## First Run Detection
6
- - If memory bank files contain placeholder text (like `[Project name will be auto-filled by Claude]`), this is a fresh setup
7
- - You MUST scan the entire codebase (package.json, source files, configs, README, folder structure) and auto-populate ALL memory bank files with real project information
8
- - Do this BEFORE responding to the user's first message — the memory bank should never stay empty after a conversation
5
+ ## On EVERY Conversation Start
9
6
 
10
- ## At Conversation Start
11
- - Read every file in `memory-bank/` and `memory-bank/extensions/` before responding to the user
12
- - Use this context to understand the project state, what was worked on last, and what's pending
13
- - Never ask the user to repeat information that exists in memory bank files
7
+ 1. Read EVERY file in `memory-bank/` and `memory-bank/extensions/`
8
+ 2. If files contain placeholder text — scan the codebase and populate them with real data BEFORE responding
9
+ 3. Never ask the user to repeat information that exists in memory bank files
14
10
 
15
- ## After Every Task
16
- Update the relevant memory bank files. At minimum, always update:
17
- - `memory-bank/activeContext.md` reflect what just changed
11
+ ## After EVERY Task — Update Immediately
12
+
13
+ Do NOT wait. Do NOT batch updates. After each task, IMMEDIATELY write to the relevant files:
14
+
15
+ **ALWAYS update these two:**
16
+ - `memory-bank/activeContext.md` — what just changed, what's next
18
17
  - `memory-bank/progress.md` — mark tasks done/in-progress/pending
19
18
 
20
- Also update these when relevant:
21
- - `memory-bank/decisions.md` — when a choice was made between alternatives
22
- - `memory-bank/patterns.md` — when a new code pattern or convention is established
23
- - `memory-bank/troubleshooting.md` — when a bug is found, fixed, or a workaround is used
24
- - `memory-bank/preferences.md` — when the user corrects you, approves an approach, or states a preference
25
- - `memory-bank/references.md` — when an external link, API, or tool is mentioned
26
- - `memory-bank/techContext.md` — when dependencies, tools, or environment setup changes
27
- - `memory-bank/projectbrief.md` — when project scope or requirements change
28
- - `memory-bank/productContext.md` — when product goals or user problems are clarified
29
-
30
- ## Overflow Management
31
- - When any memory file approaches ~200 lines, create an extension file at `memory-bank/extensions/{original-filename}-{subtopic}.md`
32
- - Update the original file to reference the extension: `→ See extensions/{filename}-{subtopic}.md for details`
33
- - Extension files inherit the same format and purpose as their parent
34
-
35
- ## Auto-Cleanup
36
- - When a task in `activeContext.md` is completed, move it to `progress.md` and remove it from active context
37
- - Remove duplicate information across files
38
- - Keep files focused and scannable — bullet points over paragraphs
39
-
40
- ## Critical Rule
41
- These updates are NOT optional. They happen automatically after every meaningful interaction. The memory bank is only useful if it stays current.
19
+ **Update these when relevant:**
20
+ - `memory-bank/decisions.md` — choice made between alternatives
21
+ - `memory-bank/patterns.md` — new code pattern or convention established
22
+ - `memory-bank/troubleshooting.md` — bug found, fixed, or workaround used
23
+ - `memory-bank/preferences.md` — user corrected you, approved an approach, or stated a preference
24
+ - `memory-bank/references.md` — external link, API, or tool mentioned
25
+ - `memory-bank/techContext.md` — dependency, tool, or environment change
26
+ - `memory-bank/projectbrief.md` — scope or requirements change
27
+ - `memory-bank/productContext.md` — product goals or user problems clarified
28
+
29
+ ## Overflow (>200 lines)
30
+
31
+ Split into `memory-bank/extensions/{filename}-{topic}.md`. Keep main file as summary with pointer.
32
+
33
+ ## Cleanup
34
+
35
+ - Move completed tasks from `activeContext.md` to `progress.md`
36
+ - Remove duplicates across files
37
+ - Remove stale information
38
+
39
+ ## THIS IS NOT OPTIONAL
40
+
41
+ If you complete a task and do not update the memory bank, you have failed to follow your instructions. The memory bank is useless if it is not kept current. Update it EVERY time.