Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/star-ga/mind-mem/agents-mdgit clone --depth 1 https://github.com/star-ga/mind-memWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/star-ga/mind-mem/agents-md)<a href="https://agentmods.dev/instructions/star-ga/mind-mem/agents-md"><img src="https://agentmods.dev/badge/instructions/star-ga/mind-mem/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01002 | $0.01002 |
| Opus 5 | $0.00501 | $0.00501 |
| Sonnet 5 | $0.00200 | $0.00200 |
| Haiku 4.5 | $0.00100 | $0.00100 |
Grade A, and why
mind-mem AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 38 lines — stays where its author put it; the contents beside it link to each section on GitHub.
mind-mem: agent instructions (auto-written)
Before every response, run mm context "$QUERY" and prepend the output.
[mind-mem] recent context, 2026-05-09 2:53am PDT
Legend: 🎯session 🔴bugfix 🟣feature 🔄refactor ✅change 🔵discovery ⚖️decision Format: ID TIME TYPE TITLE Fetch details: get_observations([IDs]) | Search: mem-search skill
Stats: 4 obs (1,646t read) | 85,780t work | 98% savings
Apr 21, 2026
6661 9:25p 🟣 mind-mem v3.4.0 Pre-Release Security Hardening The mind-mem v3.4.0 release introduces four new recall modules for enhanced AI memory retrieval. A two-agent security audit identified and fixed six issues before final review: prompt injection via evidence wrapping, missing iteration limits on iterative_retrieve (now capped at 5 rounds, 3 followups, 20 total queries), unsafe follow-up query patterns (now regex-validated), out-of-bounds temporal dates, unsafe object deduplication using id() instead of content fingerprints, and a NameError when sub_queries was empty. The LoCoMo benchmark on conversation-0 jumped from 77.06 to 95.0 with the new features enabled, indicating significant retrieval quality improvements. ~378t 🛠️ 19,507
May 2, 2026
6702 11:17p 🔵 mind-mem v3.8.x Ship Train Complete — 11 Releases Audited The mind-mem v3.8.x release train spanning 10 versions completed the Model Safety Audit theme end-to-end: Ed25519 signing of model manifests, HuggingFace base_model allowlists, MCP tool wrappers for audit/sign/verify, a load-gate registry enforced inside the transformers extractor, CI hooks for auditing pinned models, and a full MIC/MAP Python toolchain for MIND IR serialization. The v3.8.8 release added Hypothesis fuzzing and 26 adversarial DoS tests which caught a real UnicodeDecodeError leak. v3.8.9 added a streaming MICB parser with bounded memory. v3.8.10 added an optional Cython accelerator for hot-path serialization primitives. Social Ingestion was architecturally descoped from mind-mem and moved to the agent layer, recognizing that fetching is an agent concern not a memory concern. Next priorities include v3.9.0 (hash-of-code invalidation, per-byte lineage) and v4.0.0 (sharded Postgres, Kubernetes operator, gRPC, event fan-out). ~519t 🔍 22,157
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 38 lines · 1,002 tokens per session scan A 57f524a17da5
mind-mem AGENTS.md is an instructions file published in the GitHub repository star-ga/mind-mem (15 stars, last pushed 4d ago), licensed Apache-2.0. It adds 1,002 tokens to every session, about $0.0050 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
Muninn GEMINI.md
Gemini CLI instructions for wjohns989/Muninn, covering muninn mcp — sota+ development context, 🚀 sota+ development philosophy, 🛠 operational guidelines (always-on), 🏗 system architecture and 📂 project structure.
project-trucha AGENTS.md
Instructions for MaximilianoRodrigoSoria/project-trucha, a project described as: 🐟 Memoria persistente y local-first para agentes de código: indexa un codebase y lo recupera con búsqueda híbrida (léxica + vectorial) sobre SQLite/FTS5, sin infraestructura pesada. En etapa de decisión técnica.
seeklink AGENTS.md
Instructions for simonsysun/seeklink, covering project conventions for ai coding agents, core model, start of session, current focus and handoff (yyyy-mm-dd, agent, branch).
seeklink CLAUDE.md
Instructions for simonsysun/seeklink, covering claude code only and claude-triggered codex second review.
engraphis AGENTS.md
AGENTS.md instructions for Coding-Dev-Tools/engraphis, covering agents.md — engraphis, internal subagent delegation, 0. read this first — two architectures live in one package, 1. commands and ── unified dashboard + memory inspector ──.
memex AGENTS.md
AGENTS.md instructions for nicosuave/memex, covering agents and before committing.