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/modeled-information-format/mnemonicWrote 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/modeled-information-format/mnemonic/gc)<a href="https://agentmods.dev/commands/modeled-information-format/mnemonic/gc"><img src="https://agentmods.dev/badge/commands/modeled-information-format/mnemonic/gc/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/commands/modeled-information-format/mnemonic/gc"><img src="https://agentmods.dev/badge/commands/modeled-information-format/mnemonic/gc.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 | $0.00007 | $0.03067 |
| Opus 5.5 | $0.00003 | $0.01227 |
| Sonnet 5.5 | $0.00001 | $0.00613 |
| Haiku 4.5 | $0.00001 | $0.00307 |
Grade B, and why
gc scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
OLD_FILES=$(find "$HOME/.claude/mnemonic/$ORG" "./.claude/mnemonic" -name "*.memory.md" -mtime +$MAX_AGE 2>/dev/null) How it starts
The opening of the file, as written. The whole thing — 360 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Memory
Search first: /mnemonic:search {relevant_keywords}
Capture after: /mnemonic:capture {namespace} "{title}"
Run /mnemonic:list --namespaces to see available namespaces from loaded ontologies.
/mnemonic:gc
Garbage collect expired memories based on TTL and decay, with optional compression.
Arguments
--dry-run- Show what would be removed without removing--archive- Move to archive instead of deleting--force- Skip confirmation prompts--min-strength- Minimum strength threshold (default: 0.1)--max-age- Maximum age in days (default: 365)--compress- Compress large old memories before archiving--compress-only- Only compress, don't archive/delete--compress-threshold- Minimum lines for compression (default: 100)
Procedure
Step 1: Parse Arguments
DRY_RUN="${DRY_RUN:-false}"
ARCHIVE="${ARCHIVE:-true}"
FORCE="${FORCE:-false}"
MIN_STRENGTH="${MIN_STRENGTH:-0.1}"
MAX_AGE="${MAX_AGE:-365}"
ORG=$(git remote get-url origin 2>/dev/null | sed -E 's|.*[:/]([^/]+)/[^/]+\.git$|\1|' | sed 's|\.git$||')
[ -z "$ORG" ] && ORG="default"
ARCHIVE_DIR="$HOME/.claude/mnemonic/.archive/$(date +%Y-%m)"
Step 2: Find Candidates
By Age
echo "=== Finding GC Candidates ==="
echo ""
echo "Old memories (> $MAX_AGE days):"
OLD_FILES=$(find "$HOME/.claude/mnemonic/$ORG" "./.claude/mnemonic" -name "*.memory.md" -mtime +$MAX_AGE 2>/dev/null)
OLD_COUNT=$(echo "$OLD_FILES" | grep -c . 2>/dev/null || echo 0)
echo " Found: $OLD_COUNT"
By TTL Expiry
echo ""
echo "Expired TTL:"
NOW=$(date +%s)
TTL_EXPIRED=""
for f in $(find "$HOME/.claude/mnemonic/$ORG" "./.claude/mnemonic" -name "*.memory.md" 2>/dev/null); do
TTL=$(grep "^ ttl:" "$f" 2>/dev/null | sed 's/.*ttl: //')
CREATED=$(grep "^created:" "$f" 2>/dev/null | sed 's/created: //')
if [ -n "$TTL" ] && [ -n "$CREATED" ]; then
# Parse TTL (simplified - handles P{N}D format)
TTL_DAYS=$(echo "$TTL" | sed 's/P\([0-9]*\)D/\1/')
if [ -n "$TTL_DAYS" ]; then
# Calculate if expired (simplified check via file mtime)
FILE_AGE=$(( ( NOW - $(stat -f %m "$f" 2>/dev/null || stat -c %Y "$f" 2>/dev/null) ) / 86400 ))
if [ "$FILE_AGE" -gt "$TTL_DAYS" ]; then
TTL_EXPIRED="$TTL_EXPIRED $f"
echo " $f (TTL: ${TTL_DAYS}d, Age: ${FILE_AGE}d)"
fi
fi
fi
done
TTL_COUNT=$(echo "$TTL_EXPIRED" | wc -w | tr -d ' ')
echo " Total: $TTL_COUNT"
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.
- 9d ago Changed 28ca8d0310a1
- 10d ago Changed · +1 lines 0d8bcd1db595
- 1mo ago First seen · 359 lines · 7 tokens per session scan B b0223f8ba1f2
gc is a command published in the GitHub repository modeled-information-format/mnemonic (23 stars, last pushed 6d ago), licensed MIT. It adds 7 tokens to every session and 3,067 once invoked, about $0.0000 per session on Opus 5.5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
dream-defrag
A command for pruning Claude Code’s saved memory files, including MEMORY.md and other memory notes.
context-preload
A command for managing and setting up ContextPreloader.
voice-refresh
Refresh the founder-voice DNA from new Granola meetings (interactive, MCP-backed). Pulls meetings since the last refresh, harvests Assaf-only speech, runs the Stage 2-3 orchestrator, and presents a founder-gated delta. Never auto-writes voice-dna.md.
q-research
Read the research-mode skill's SKILL.md for the full ruleset before proceeding. Follow all constraints, the source lookup cascade, the token budget, and the "what counts as cited" rules exactly.
status
Show memex statistics and status including projects, memos, and pending items.
learn
Command "learn" from datacore-one/datacore, covering /learn, purpose, when to use, usage and quick capture.