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 rekal-dev/rekal-cli --skill rekalgit clone --depth 1 https://github.com/rekal-dev/rekal-cliWrote 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/rekal-dev/rekal-cli/rekal)<a href="https://agentmods.dev/skills/rekal-dev/rekal-cli/rekal"><img src="https://agentmods.dev/badge/skills/rekal-dev/rekal-cli/rekal/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/rekal-dev/rekal-cli/rekal"><img src="https://agentmods.dev/badge/skills/rekal-dev/rekal-cli/rekal.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00091 | $0.02069 |
| Opus 5.5 | $0.00036 | $0.00828 |
| Sonnet 5.5 | $0.00018 | $0.00414 |
| Haiku 4.5 | $0.00009 | $0.00207 |
Grade A, and why
rekal 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 1mo 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 — 146 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Rekal — which substrate answers this?
Most wasted effort is the wrong substrate. Decide before you grep or recall.
| Substrate | Holds | Tense | Reach it with | Answers |
|---|---|---|---|---|
| Tree | current code | now | grep / read | what does X do, where is it |
| Knowledge | current prose | now | rekal "<q>" → Read HEAD |
convention / what we know |
| Ledger | session intent | past | rekal "<q>", drill rekal query --session |
why, tried, rejected |
| Map | structure | — | map.sh + workflow |
how is it built |
Boundary
- Is it true now, or something that was?
- Was — a reason, a rejected path, a past correction, or a fact whose only record is a past conversation → Ledger. (On a pure-dialogue corpus only the ledger has content — go there.)
- If now — code or prose?
- Code (path/symbol, present tense) → Tree. Grep; do not recall.
- Prose → Knowledge. Never invent an episode when HEAD prose answers.
grep for code that is · knowledge for prose that is · ledger for the why that was.
Commands — text by default, --json for machines
Rekal's read commands print compact agent-readable text; add --json only when
a program needs to parse it.
rekal "<q>"— recall. Prints a seed digest: line 1 is the verdict (INJECT/KNOWLEDGE/SILENCE), then per-seedsid conf=… t<n> "snippet". One call already widens itself — it fuses several deterministic reformulations of your query (keyword-only, clause splits, a temporal variant) so you get the full picture in one go; still reformulate by hand only when the answer needs a genuinely different angle the mechanical variants miss. A seed may carry[reached N× drilled M×· "past query"]before its snippet — a usage hint.reachedcounts how often the search surfaced it, which is the engine quoting itself: on a small store nearly everything is reached, so a bare high count means little.drilledcounts how often an agent opened it — that is the load-bearing signal and a good first drill. The echoed query is the one that most often surfaced this memory, so it shows how the need is usually framed. Neither raisesconf=— judge relevance fromconf=+ content as always. No tag just means newly surfaced, not worse.rekal find "<term>" [role]— every ledger mention of a term, complete and in time order (the "all / every / how many" sweep). A partial list is a wrong answer to a set question — this is the set.rekal query --session <sid|ulid> [--offset N --limit 5 --role …]— drill a session into readable turns.rekal query --sql "SELECT …"for analytical / complete-set SQL (seereferences/reference.mdfor the full schema;tsis a TIMESTAMP — useBETWEEN, notLIKE).
What ships with it
11 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/ledger.md 18 KB
- references/map.md 1.5 KB
- references/reference.md 3.6 KB
- references/wiki.md 1.0 KB
- references/workflows/complete-set.md 1008 B
- references/workflows/duration.md 1.1 KB
- references/workflows/event-time.md 1.2 KB
- references/workflows/inference.md 1.0 KB
- references/workflows/point-fact.md 1.2 KB
- scripts/map.sh 2.4 KB runs code
- scripts/wiki-gate.sh 1018 B runs code
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.
- 1mo ago First seen · 146 lines · 91 tokens per session scan A c3774d215dc2
rekal is a skill published in the GitHub repository rekal-dev/rekal-cli (207 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 91 tokens to every session and 2,069 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 2026-08-30.
Other skills, from other repositories
context-engineering
Use when designing, reviewing, or debugging how an agent's context window gets filled, pruned, or shared — choosing what loads at boot versus on demand, sizing an install or an always-loaded file, fixing an agent that drifts, repeats itself, or forgets constraints mid-task, planning compaction or summarization…
claude-compaction-restore
Use when a Claude Code session is about to compact, has just compacted, reached its context limit, resumed after /compact, or must rebuild its working picture of a long-running seat from its own restore map, its JSONL transcript and the files it touched.
retiring-and-inheriting-a-seat
Use when you are a sitting agent near your context threshold (85%) and a PLANNED seat transition is due — retire deliberately and hand your seat to a fresh successor primed from a packet, rather than let compaction degrade you. Covers the handover packet, the append-only lineage ledger (one row per tenure)…
session-compaction-and-restore
Use when reasoning about what survives compaction/context-loss/restart/seat-refresh, designing a high-fidelity restore packet, or distinguishing native runtime resume vs fork vs artifact-backed mental-model rebuild. Covers the 4 failure modes that prevent honest restore (compacted seat drops hot potato; restore packet…
mnemon
Persistent memory CLI for LLM agents. Store facts, recall past knowledge, link related memories, manage lifecycle.
mnemon
Persistent graph-based memory. Recall context before responding, remember insights after. Each group has private memory; global memory is read-only.