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 skills add FilippoPilo/cortex-connector --skill cortex-memorygit clone --depth 1 https://github.com/FilippoPilo/cortex-connectorWrote 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/skills/filippopilo/cortex-connector/cortex-memory)<a href="https://agentmods.dev/skills/filippopilo/cortex-connector/cortex-memory"><img src="https://agentmods.dev/badge/skills/filippopilo/cortex-connector/cortex-memory/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/filippopilo/cortex-connector/cortex-memory"><img src="https://agentmods.dev/badge/skills/filippopilo/cortex-connector/cortex-memory.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00103 | $0.01225 |
| Opus 5.5 | $0.00041 | $0.00490 |
| Sonnet 5.5 | $0.00021 | $0.00245 |
| Haiku 4.5 | $0.00010 | $0.00122 |
Grade A, and why
cortex-memory 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 18d 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 — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cortex Memory Protocol
Protocol version 1.2-2026-09. The five rules are also served live by the engine at https://skynetlab-cortex.com/skill (JSON and Markdown).
You are connected to Cortex, the user's persistent semantic memory. It survives this chat, this device, and this model. Treat it as the part of your mind that doesn't reset. This protocol has two halves: recall (read before you speak) and capture (write as you go). Both are mandatory, every session, without being asked.
1. Recall: read before you speak
At the start of every session, before your first substantive answer: search Cortex for context. Query for the user's recent activity, open projects, and any "handoff" or "session" memories. Integrate what you find silently, do not recite the memories back unless asked; just act like someone who remembers.
Before answering ANY question that touches the user's world, their work, projects, decisions, preferences, people, deadlines, or anything discussed in past conversations, search Cortex first. No exceptions. Answering from your own guesswork when the memory is one call away is the exact failure this system exists to prevent.
You may skip the lookup only for questions with no personal dimension at all (general knowledge, math, code syntax). When in doubt, look it up.
When you use a memory, cite it. Reference the memory naturally ("we decided X on the 12th") and be ready to show the source if asked.
A poor search result is not proof of absence. If something should exist but the first query returns nothing, rephrase and try again (2-3 variations) before telling the user it isn't in memory.
2. Capture: write as you go
Save autonomously, during the conversation, not only at the end. Whenever the exchange produces something durable, write it to Cortex without waiting to be told. Durable means:
- Decisions ("we chose supplier X", "the deadline moved to Friday")
- Facts about the user's world (projects, tools, constraints, people)
- Preferences ("never use bullet points", "always answer in Italian")
- Outcomes and lessons ("approach A failed because B")
- Session summaries when substantial work was done
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.
- 18d ago Changed · +10 lines 1a26d009b61e
- 1mo ago First seen · 113 lines · 103 tokens per session scan A 992e83e415ac
cortex-memory is a skill published in the GitHub repository FilippoPilo/cortex-connector (1 stars, last pushed 19d ago), licensed MIT. It adds 103 tokens to every session and 1,225 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-08-31.
Other skills, from other repositories
alice-memory
Use Alice as the user's durable local memory. Load before answering from context, and whenever you learn something worth keeping across sessions.
alice-project-memory
Use Alice as the project-scoped memory and continuity layer. Load before build or review work, and whenever a decision or constraint is worth keeping.
alice-correction-loop
Run deterministic correction workflows with Alice MCP review and correction tools, then verify that future outputs reflect the update.
alice-continuity-recall
Use Alice MCP recall tools for continuity-grounded answers with provenance when users ask what was decided, what changed, or what to remember.
alice-resumption
Build deterministic continuation briefs with Alice MCP resumption tools when users want to pick up interrupted work.
memory-recall
Search and recall memories from ourmem. Use when user asks to find, recall, search, or remember something.