Forget — delete agent memory

Trust Score 74/100

Guides safe deletion of specific observations or sessions from agentmemory: search, confirm with user, then call memory_governance_delete with explicit memory I

triggers:forget thisdelete memoryremove observationmemory deletepurge memoryforget sessionremove user data

GitHub SKILL.md

What it does

This skill lets an agent safely remove specific observations or sessions from the agentmemory store. It prescribes a non-destructive workflow: search for matching memories, present results to the user, request explicit confirmation, then delete by memory IDs and report the deletion count.

When to use it

Use when a user explicitly requests data removal (phrases like "forget this", "delete memory") or when sensitive information must be purged for privacy compliance. Always require explicit user confirmation before performing deletions.

What's included

Scripts: none bundled (has_scripts=false).
References: relies on MCP tools (memory_smart_search, memory_governance_delete) made available by the agentmemory server; instructions cover required MCP calls and fallback steps if tools are unavailable.
Instructions: step-by-step: search (limit 20), display session/observation IDs, ask for confirmation, call memory_governance_delete with memoryIds and optional reason, then confirm deletion count.

Compatible agents

Designed for agents that integrate MCP plugins (Claude Code, Codex/Claude-style MCP-enabled runtimes) and any agent with access to the agentmemory MCP interface.

Audit Summary

A skill that guides an agent to delete specific observations or sessions from agentmemory via MCP tools. Requires explicit user confirmation before any deletion, which is good safety practice. No bundled scripts — purely instructional via SKILL.md. The instructions are clear but narrowly scoped to agentmemory's specific MCP tools (memory_smart_search, memory_governance_delete), limiting portability. Falls back gracefully with troubleshooting steps if MCP tools aren't available.

Watch Out

Only works with agentmemory MCP server — not portable to other memory systems
No $ARGUMENTS validation — empty or malformed input goes straight to search
Deletes by memory ID only; dropping a full session requires collecting all IDs manually

Notes

Destructive operation skill that appropriately requires user confirmation. No security red flags — no remote code execution, credential exposure, or exfiltration risk. Minor deduction for no input validation and tight coupling to a single MCP server's API. Architecture is basic but functional — frontmatter present, no scripts/references separation. Usefulness is niche since it only serves agentmemory users.

Information

Repository: agentmemory
Stars: 1,838
Installs: 0

Trust Score

Overall74

Security88

Code Quality62

Architecture55

Usefulness52

Related Skills

Retriever (Membrane CLI)

Integrate Retriever search and discovery via the Membrane CLI — manage connections and run queries across Retriever data sources.

AWP (Agent Work Protocol)

Tooling and scripts for onboarding, staking, allocation, and managing agents on the AWP network (Base/Ethereum/Arbitrum/BSC). Includes safe, opt-in daemon and r

Caveman Compress

Compress natural-language memory files into a compact 'caveman' format to save LLM tokens while preserving code, links, and structure.

ClawPod / Massive Unblocker

Bypass anti-bot restrictions and fetch rendered HTML or structured search results via Massive's Unblocker API (handles CAPTCHAs, JS rendering, geo-restrictions)

Retriever (Membrane) Integration

Integrate with Retriever via the Membrane CLI to search, query, and proxy Retriever data (documents, snippets, queries) with built-in auth and pagination.

Ip2Location IO Automation (Composio)

Automate Ip2location IO tasks via Composio's Rube MCP toolkit — discover tools, verify connections, and execute schema-compliant workflows safely.

GRC & Compliance (ServiceNow)

Guides Governance, Risk & Compliance (GRC) work in ServiceNow: policies, controls, risk records, audits and ES5 script examples for automating GRC workflows.

Smalltalk MCP

Connects an LLM (Claude/OpenAI/etc.) to a live Smalltalk (Squeak/Cuis) image over MCP for browsing, evaluating and modifying Smalltalk code.

Back to Skills