skills.rest
vesviet
vesvietCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-semantic-memory

Read and write persistent memory entries so agents retain codebase facts across conversations.

Updated Feb 10, 2026
Search Tags:#knowledge graph#context engineering#vector search#agent memory#semantic memory#episodic memory#secret scanning
One-click install
npx skills add https://github.com/vesviet/agent-skills --skill agent-semantic-memory-vesviet
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-semantic-memory
Source: https://github.com/vesviet/agent-skills/tree/main/core/skills/agent/agent-semantic-memory
Command: npx skills add https://github.com/vesviet/agent-skills --skill agent-semantic-memory-vesviet

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Agents lose all knowledge of past fixes, codebase conventions, and architectural facts when a conversation ends, forcing them to rediscover the same patterns and repeat the same mistakes. This Skill provides a disciplined process for persisting and retrieving reusable knowledge across sessions. ## Core Features & Use Cases - Tiered Memory Model: Distinguishes working, episodic (incident log), and semantic (temporal knowledge graph) memory with concrete YAML entry formats for each. - Zero Secret Ingestion: Enforces regex and entropy secret scanning on every write payload so credentials, tokens, and PII never reach the memory store. - Staleness Validation: Requires last_verified timestamps and cross-checking retrieved memories against the live codebase before acting on them. - Use Case: After fixing a non-obvious Postgres JSON query bug in a Laravel repo, write a tagged episodic entry; weeks later, a new session retrieves it before touching the same query layer and avoids repeating the debugging session. ## Quick Start Use the agent-semantic-memory skill to record the bug fix we just made as a tagged episodic memory entry with a last_verified timestamp.

Frequently Asked Questions about agent-semantic-memory

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I give an AI agent persistent memory across conversations?▼

Store reusable knowledge as tagged episodic or semantic entries in a persistent memory store, then query it before starting work in a known repo. Each entry needs repo context, type tags, and an ISO-8601 last_verified timestamp so retrieval can be filtered and validated.

What is the difference between episodic and semantic memory for agents?▼

Episodic memory logs past actions and outcomes, such as how a specific bug was fixed. Semantic memory stores durable facts about codebases and architectures, like which ORM a repo uses, typically as vector embeddings plus entity-relation triples.

How do I prevent secrets from being written to agent memory?▼

Run regex and entropy secret scanners on every write payload before persistence, and reject any entry containing credential patterns. Never store tokens, passwords, customer PII, or private API keys in memory entries.

Can agents trust retrieved memories without verification?▼

No. Retrieved memories are untrusted hypotheses that may be stale or poisoned. Check the last_verified timestamp, cross-reference against the live codebase before acting, and never let a memory entry override the active user request.

When should an agent write to long-term memory?▼

Write only when knowledge is reusable across future tasks: after fixing a non-obvious bug, discovering a codebase convention, or when the same mistake has occurred more than once. Entries must be specific and actionable, not vague summaries.

Related Skills

Explore More Software Engineering Skills→
bykoleksii-hardobykoleksii-hardo

memory-systems

Integrate persistent semantic memory with vector stores and knowledge graphs.

Community
Advanced
QuixiAIQuixiAI
609

core-memory

Recalls, stores, and searches persistent memories across agent sessions.

Official
Advanced
melikhanmutlumelikhanmutlu

agent-memory-systems

Structure agent memory with short-term, long-term, and retrieval strategies.

Community
Advanced
arthur0narthur0n

graphify

Query a codebase knowledge graph for architecture and file-relationship questions via the graphify CLI.

Community
Basic