memory-gate
Skill0 starsChanged 45 days agoMITView on GitHub
Run generalizability gate after fixing non-trivial bugs or making architectural decisions. Evaluates whether lessons should be promoted from project-local to cross-project global memory. USE WHEN: A bug fix involves a named tool/library/pattern, or an architectural decision applies beyond this project.
In this repository10 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- memory-retrospectiveSkill
- phase-distillationSkill
- uws-kbSkill
- uws-orchestrateSkill
- uws-research-leadSkill
- vector-memorySkill
- workflow-checkpointSkill
- workflow-recoverySkill
- workflow-statusSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
After storing a bug-fix or architecture memory to local DB, evaluate:
- Distinctive: line 33
Call: mcp__vector_memory_global__store_memory(content, category, tags)
≈ “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
Generalizability Gate
After storing a bug-fix or architecture memory to local DB, evaluate:
Q1: Does the root cause involve a named tool, library, or architectural pattern (not just this project's code)? Q2: Could someone hit this exact issue in a different project? Q3: Can you state the lesson in one sentence WITHOUT referencing any file path, variable name, or project-specific term from THIS project?
ALL THREE = YES:
Compose abstracted lesson (NO project file paths, NO project variable names):
content="<TOOL_OR_PATTERN>: <root cause mechanism>
FIX: <fix approach> APPLIES_TO: <scope>"
Select server category using global mapping:
anti-pattern -> architecture | tool-gotcha -> bug-fix |
design-lesson -> architecture | library-compat -> other |
workflow-improvement -> other
Embed plan category in content: "CATEGORY: <plan-category> | ..."
Select tags: [<root-cause-tag>, <scope-tag>, <fix-pattern-tag>] (NO colons)
Call: mcp__vector_memory_global__store_memory(content, category, tags)
ANY = NO: Local store only. No global promotion.
CALIBRATION: After every 10 global promotions, run adversarial review:
"Which of the last 10 global stores are actually project-specific?"
For false promotions, store corrective entry:
"SUPERSEDES: <summary>. REASON: project-specific."
(No delete_memory tool -- supersede pattern only.)
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.