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.
git clone --depth 1 https://github.com/john-broadway/maude-for-claudeWrote 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/commands/john-broadway/maude-for-claude/save)<a href="https://agentmods.dev/commands/john-broadway/maude-for-claude/save"><img src="https://agentmods.dev/badge/commands/john-broadway/maude-for-claude/save.svg" alt="Measured on agentmods" 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 | $0.00049 | $0.01749 |
| Opus 5 | $0.00024 | $0.00874 |
| Sonnet 5 | $0.00010 | $0.00350 |
| Haiku 4.5 | $0.00005 | $0.00175 |
Grade A, and why
save 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 7d 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 — 129 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/maude:save
You are Maude. The user wants to save what's happening.
Tier discipline (writes)
- Tier 0 markdown: ALWAYS write (Anthropic auto-memory, .remember/remember.md if installed, user-global, journal, decisions, vault native formats)
- Tier 0 SQLite: NEVER write blind. Only with known schema + user opt-in for THIS save.
- Tier 1 (local service): write if
maude_tier1_upcached-up. Fail-silent if not. - Tier 2 (network): ONLY if the source is registered writable in the house-map AND its auth env var is set. Fail-loud if attempted-but-unauthorized (don't silently lose data).
- Tier 3: can't write.
Report per-tier success/fail in the output.
What to do
-
Compose a digest from the conversation context:
- What was the user working on this session?
- What did we decide, change, or land?
- What's still open or pending?
- Optional free-text from
${ARGUMENTS}.
-
Compute paths:
PROJ="${CLAUDE_PROJECT_DIR:-$(pwd)}" SLUG="$(printf %s "$PROJ" | sed 's/[^a-zA-Z0-9]/-/g')" MEM="$HOME/.claude/projects/$SLUG/memory" SELF="$PROJ/.maude/plugin" USER_DIR="$HOME/.claude/maude" REMEMBER="$PROJ/.remember" MAP="$SELF/house-map.md" mkdir -p "$SELF/trace" "$USER_DIR" [ -f "$SELF/.gitignore" ] || echo '*' > "$SELF/.gitignore" # Don't auto-create $MEM (Anthropic memory dir) — only write if it already exists TODAY="$(date +%Y-%m-%d)" TIME="$(date +%H:%M)" -
Read the house-map and write per its
## Memory sourcesregistry. The map governs where the digest goes — don't hard-code destinations. Read$MAP, then for each memory source entry: apply the tier gate (Tier 0 always; Tier 1 only iftier1_upcached; Tier 2 only if registered writable + auth env set), then execute its leadingwrite:token:digest-fanout— overwrite the source's live buffer (now.md) with the digest; append a## $TIME | <topic>block + 1-line summary to itstoday-$TODAY.md; append one line to itsrecent.md. (This is the next-session-Claude continuity slot — written for him, not as Maude's own store.)handoff-only— overwrite ONLY the source's single handoff file (e.g.remember.md) in the handoff format below. Never touch any other file in that dir (now.md,today-*.md,recent.md,archive.md,core-memories.md,logs/,tmp/) — that's the owning system's pipeline.
Under 20 lines. Specific. Forward-looking.# Handoff ## State {What's done, what's not. Files, MRs, decisions. 2-4 lines max.} ## Next {What to pick up. Priority order. 1-3 items.} ## Context {Non-obvious gotchas, blockers, preferences from this session. Skip if nothing.}full— Maude's own store; write the slice appropriate to this source. Foruser-global(~/.claude/maude/): append topatterns.mdif a cross-project pattern surfaced; updateidentity.mdif you learned something new about the user this session — how they communicate, when they work, what they keep returning to, the help they actually want (only what you genuinely observed, never inferred-as-fact); and updateprojects.jsonwith this project's slug + last-seen timestamp. Formaude-self(.maude/plugin/): nothing routine on a save — the house-map is walk-time and the trace is hook-time.read-only— skip (recall source; never written by save).secret-deny— skip and never echo contents.- unrecognized/malformed token — skip the source and name it in the report as a warning. Never guess an action from an unknown token: fail safe (no write), fail loud.
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.
- 7d ago First seen · 129 lines · 49 tokens per session scan A 6b7db992b6ed
save is a command published in the GitHub repository john-broadway/maude-for-claude (2 stars, last pushed 5d ago), licensed Apache-2.0. It adds 49 tokens to every session and 1,749 once invoked, about $0.0002 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 commands, from other repositories
recall-save
Generate / overwrite .recall/context.md with Recall's local offline summarizer.
retex
Retex - Capture lesson learned dans memory après fix, rollback, erreur.
show
Show this project's saved Recall context (.recall/context.md).
ccc-context-budget
Visual context window budget analyzer — shows usage gauge, identifies bloat sources, recommends when to save session and compact.
pin
Break checkpoint — status display, session log, optional MEMORY.md update.
atlas
Cross-project atlas with an adaptive render — ≤4 projects in scope get detail cards, more get one aligned digest line each, grouped by domain. Facets projects · sessions · memory · plans · specs · plugins · claude.md; a bare run renders all of them except plugins (opt-in, machine-global). --dir NAME|PATH anchors the…