Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add ManceRayder42/jarvis-os/plugin install jarvis-osWrote 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/mancerayder42/jarvis-os/memory-consolidation)<a href="https://agentmods.dev/skills/mancerayder42/jarvis-os/memory-consolidation"><img src="https://agentmods.dev/badge/skills/mancerayder42/jarvis-os/memory-consolidation/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/mancerayder42/jarvis-os/memory-consolidation"><img src="https://agentmods.dev/badge/skills/mancerayder42/jarvis-os/memory-consolidation.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 | $0.00104 | $0.03522 |
| Opus 5 | $0.00052 | $0.01761 |
| Sonnet 5 | $0.00021 | $0.00704 |
| Haiku 4.5 | $0.00010 | $0.00352 |
Grade A, and why
memory-consolidation 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 11d 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 — 228 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Memory Consolidation
You are a memory-consolidation agent. Left to a manual trigger, you consolidate the window the invoking prompt gives you. If the user has wired up the opt-in scheduler (scripts/memory-tick.mjs, see references/scheduling.md), that scheduler runs on a tick that fires at most once per calendar day, whenever the machine is on and online, and derives the window from the last successful run — which may span several days if runs were missed.
Hard caps (DO NOT EXCEED)
- Max 40K input tokens total (after filtering)
- Max 5K output tokens across all 3 file updates
- Skip run entirely if fewer than 10 user messages in the window (scale this floor if the window covers more than one day — see Scope below)
- Total budget: stop and write partial if exceeded
Conversation log location
~/.claude/projects/*/*.jsonl
Claude Code writes session logs into per-cwd project directories under ~/.claude/projects/. This skill scans all of them — every project the user has worked in with Claude Code, not just one — because there's no reliable way to guess which single directory is "the" projects root on someone else's machine. If the user wants to scope this to a subset, they can set JARVIS_CONSOLIDATION_SCOPE to a narrower glob before invoking; absent that, scan everything.
Each .jsonl file is one conversation session. Lines are JSON objects with type, message, etc.
Assert before proceeding: count the matched files. If the count is 0, do NOT take the sparse-window path — that means the glob itself is broken (wrong home directory, no Claude Code history yet), which is a different problem than "a quiet stretch." Report it plainly and exit non-zero rather than writing a misleadingly cheerful "quiet day" note.
Filter: only keep messages where type === "user" or type === "assistant" AND message.content is text (not tool use). Drop all tool_use and tool_result rows to save budget. Also drop system reminders.
Also drop injected skill/system text from "user" rows. A type: "user" row is not necessarily something the user typed — the harness injects skill bodies, hook output and command output as user turns. Discard any user-turn content that starts with Base directory for this skill:, <command-name>, <local-command-, <system-reminder>, or Caveat:. Without this filter one session's apparent volume inflates ~10x and the pattern counts become meaningless.
What ships with it
4 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.
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.
- 11d ago First seen · 228 lines · 104 tokens per session scan A 9e9277ba8442
memory-consolidation is a skill published in the GitHub repository ManceRayder42/jarvis-os (5 stars, last pushed 5d ago), licensed MIT. It adds 104 tokens to every session and 3,522 once invoked, about $0.0005 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 2026-08-31.
Other skills, from other repositories
memory-to-skill
Turn workflows from your MemSearch memory into reusable skills. Use when the user asks to make/create/extract/distill a skill from what they just did or from past work, review skill candidates, install a distilled skill, or 'turn this into a skill'. Manages MemSearch procedural-memory candidates under…
learn
Must be used near the end of any non-trivial turn that produced potentially reusable tools, guidance, errors, workarounds, or workflows, so those lessons are saved for future turns.
recall
Must be used at the start of any non-trivial task involving code changes, debugging, repo exploration, file inspection, or environment/tooling investigation to surface stored guidance before analysis or tool use.
adapt-memory
Mirror a just-saved native memory into the shared evolve store so it becomes shareable and auditable.
team-memory-protocol
A shared-memory process for teams of coding agents working on the same build. It stores project knowledge, shared context, and task progress in separate files or repositories.
setup-brain
Set up or upgrade an AI-powered Obsidian vault. Interviews you, builds your vault structure (or works with what you already have), creates your CLAUDE.md memory file, installs tools, and gets you journaling — all in one conversation. Also has a repair/upgrade path for existing users.