Engram

Persistent semantic memory for AI agents — local, fast, free. Use when agent needs to recall past decisions, store new facts/preferences, search conversation history, or maintain context across sessions.

Security Scan

The skill's requirements and runtime instructions align with a local CLI-based memory tool; nothing requested is disproportionate to that purpose, but you should review the npm package before installing because CLI packages can execute code and access files.

Purpose & Capability

Name/description (local persistent semantic memory) match the declared binary 'engram' and the npm package 'engram-memory' that provides that CLI. Required binaries and declared install step are coherent with the skill's stated purpose.

Instruction Scope

SKILL.md contains only CLI usage for search/add/ingest/relate/export/import and references local files (e.g., export/import) and stdin ingestion — all expected for a local memory CLI. It does not instruct reading unrelated system files or requiring unrelated env vars.

Install Mechanism

Install uses a public npm package (engram-memory) which is a typical distribution method for a CLI. This is proportionate, but npm packages can execute arbitrary install scripts and the installed binary can perform network I/O; the skill claims 'No cloud' but that cannot be verified without inspecting the package.

Credentials

No environment variables, credentials, or config paths are requested. That is consistent with the 'local, no API keys' claim. Because no secrets are requested, there is no immediate credential overreach.

Persistence & Privilege

always:false and normal autonomous invocation are used. The skill does not request system-wide config changes or other skills' credentials. It will read/write local files (backup.json etc.) as expected for a local memory tool.

Assessment

This skill appears internally consistent for a local CLI-based memory tool, but take these precautions before installing: 1) Inspect the npm package (engram-memory) on the npm registry or its repository — check package.json for postinstall scripts and examine the binary's source if available. 2) Prefer installing in a sandbox/container or with restricted permissions to confirm it behaves offline if you require 'no cloud' guarantees. 3) Be aware that export/import commands create/read local files (backup.json) — avoid exporting sensitive data to unsecured locations. 4) Run npm audit / malware scans and verify the package author and recent activity. If you cannot review the package source, treat the install as higher risk.

Engram — Agent Memory

Local semantic memory with biological decay, typed memories, and relationship graphs. No API keys. No cloud.

Boot Sequence

engram search "<current task or context>" --limit 10

Always recall before working. Accessed memories get salience-boosted.

Storing

engram add "Client uses React with TypeScript" --type fact --tags react,client
engram add "We decided to pause ads" --type decision --tags ads
echo "Raw conversation text" | engram ingest

Types: fact, decision, preference, event, relationship

Searching

engram search "what tech stack"
engram search "pricing decisions" --type decision
engram search "client status" --agent client-agent

Relationships

engram relate <src> <tgt> --type supports
engram auto-relate <id>
engram relations <id>

Types: related_to, supports, contradicts, caused_by, supersedes, part_of, references

Key Concepts

• Decay: Unused memories lose salience daily. Recalled ones get boosted.
• Types: Filter by fact, decision, preference, event, relationship.
• Scoping: global, agent, private, shared.
• Dedup: >92% similarity auto-merges.

Quick Reference

engram stats
engram recall --limit 10
engram export > backup.json
engram import backup.json