compact

A command that reduces the amount of recent information kept in the total-recall memory system by moving useful entries to longer-term storage or removing them.

In plain words
What is it for?
Use it to review and compact total-recall's recent memory, preserve selected entries in the warm tier, and discard entries that are no longer useful.
Why use it?
It helps keep recent memory manageable and avoids sending old or unnecessary entries into the active context. You can choose a model-based review or a faster rule-based sweep.

Command

Install

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.

agentmods
npx agentmods add commands/strvmarv/total-recall/compact
Clone the repo
git clone --depth 1 https://github.com/strvmarv/total-recall
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 256 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00256
Opus 5 $0.00000 $0.00128
Sonnet 5 $0.00000 $0.00051
Haiku 4.5 $0.00000 $0.00026

Measured 2d ago against content hash fe0d9ce1d0f7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

compact 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 2d 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.

skills/commands/compact.md · 20 lines

What it actually says

/total-recall:commands compact

Compact the hot tier into the warm tier. Two modes:

Default — model-driven (LLM quality)

  1. Call the total-recall session_context MCP tool to get current hot tier entries (pinned-tier entries never appear — immune to compaction by construction).
  2. If there are 2+ hot entries, launch the total-recall:compactor agent with the entries as input, then execute its decisions:
    • carry_forward: leave in hot tier (no action)
    • promote with summary: call memory_store with the summary in warm tier, then memory_delete the source entries
    • promote without summary: call memory_promote for each entry to warm tier
    • discard: call memory_delete with the reason
  3. Report a one-line summary of what was promoted/discarded.

--fast — heuristic (no LLM)

Run the deterministic decay-based sweep instead: invoke total-recall compact --run (promotes hot entries whose decay score falls below compaction.warm_threshold). Use when you want a quick, cheap sweep without LLM summarization.

Changes

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.

  1. 2d ago First seen · 20 lines · 0 tokens per session scan A fe0d9ce1d0f7

Subscribe to this mod's changes

compact is a command published in the GitHub repository strvmarv/total-recall (14 stars, last pushed 4d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 256 tokens. 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.