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 skills/sandst1/remind/remind-curatenpx skills add sandst1/remind --skill remind-curategit clone --depth 1 https://github.com/sandst1/remindWhat 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 | $0.00081 | $0.02495 |
| Opus 5 | $0.00041 | $0.01247 |
| Sonnet 5 | $0.00016 | $0.00499 |
| Haiku 4.5 | $0.00008 | $0.00249 |
Grade A, and why
remind-curate 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 yesterday.
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 — 273 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Remind - Curating Memory
Keep the memory layer trustworthy: process pending episodes into concepts, triage contradictions, correct or retire stale items, and keep labels consistent.
Curation is where your reasoning turns raw episodes into lasting knowledge. Remind stores facts deterministically but relies on you to form patterns, identify relationships, and resolve conflicts.
When to curate
- Session end — process what accumulated during the session
- After bulk document ingestion — always run curation after reading multiple documents from different sources or time periods; this is when silent contradictions are most likely
- When
recallorsnapshot conflictsshows open conflicts - When the user corrects something already stored
The curation loop
At session boundaries or when cleaning up memory:
1. Read current state
remind snapshot pending conflicts health
This returns JSON with:
- pending.episodes: Unprocessed episodes with their entities
- conflicts.conflicts: Open conflicts with full fact details
- health: Summary of issues needing attention (pending count, open conflicts, orphan concepts)
2. Find contradictions before forming concepts
This step is mandatory after bulk document ingestion. For each major theme in the pending episodes, run a recall query to surface what is already in the store:
remind recall "structural system framing" -k 8
remind recall "heating mechanical system" -k 8
remind recall "budget cost schedule" -k 8
Compare the recall output against the pending episodes. Look for:
- The same attribute stated with different values (e.g., two roof pitches, two budgets)
- A pending episode that explicitly supersedes an earlier document's claim
- Nearby episode IDs that were flagged during ingestion (shown in
rememberoutput as "Nearby — review for conflicts")
When you find a contradiction, record it immediately — do not defer to later:
remind apply << 'EOF'
conflict a=ep:<older_id> b=ep:<newer_id> note="doc A says timber frame, doc B says steel moment frame — needs triage"
EOF
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.
- yesterday First seen · 273 lines · 81 tokens per session scan A 9573d4f4b66d
remind-curate is a skill published in the GitHub repository sandst1/remind (83 stars, last pushed 22d ago), licensed Apache-2.0. It adds 81 tokens to every session and 2,495 once invoked, about $0.0004 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 skills, from other repositories
codeguard
A CodeGuard security skill that helps AI coding agents write secure code and prevent common vulnerabilities. Use this skill when writing, reviewing, or modifying code to ensure secure-by-default practices are followed.
memory-safe-migration
Guide secure migration of code from memory-unsafe languages (C, C++, Assembly) to memory-safe languages (Rust, Go, Java, C#, Swift). Use when migrating or rewriting legacy C/C++ code, designing FFI boundaries between safe and unsafe code, writing new modules in existing C/C++ codebases, reviewing mixed-language…
CodeGuard MCP Meta Skill
Instructs AI coding agents to invoke CodeGuard MCP Server security rules before writing or reviewing code.
security-review
Comprehensive security code review workflow for a target repository, producing a markdown report with findings and recommendations.
nocturnusai-admin
Use when managing NocturnusAI databases, tenants, health checks, metrics, backups, API key management, RBAC configuration, admin operations, or operational monitoring. Triggers on: NocturnusAI database, tenant, health, metrics, backup, API key, RBAC, admin, monitoring, operational.
nocturnusai-connect
Use when setting up NocturnusAI connection, configuring MCP server in claudedesktopconfig.json or .mcp.json, setting up API keys, auth, RBAC bootstrap, creating databases/tenants, or troubleshooting connection issues. Triggers on: setup, connect, configure, MCP, auth, API key, tenant, database, bootstrap, NocturnusAI.