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 agentmods add skills/aendrix03/graft/memory-auditnpx skills add AEndrix03/Graft --skill memory-auditgit clone --depth 1 https://github.com/AEndrix03/GraftWrote 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/aendrix03/graft/memory-audit)<a href="https://agentmods.dev/skills/aendrix03/graft/memory-audit"><img src="https://agentmods.dev/badge/skills/aendrix03/graft/memory-audit.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 | $0.00125 | $0.01566 |
| Opus 5 | $0.00063 | $0.00783 |
| Sonnet 5 | $0.00025 | $0.00313 |
| Haiku 4.5 | $0.00013 | $0.00157 |
Grade A, and why
memory-audit 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 4d 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 — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.
memory-audit — Read-only health check + actionable suggestions
The graph rots like any knowledge base: people save sloppy summaries, save the same fact with three different keywords, save things that never get reused. This skill produces a single readable report and a punch-list of concrete maintenance actions for the user to approve.
Never modify the graph from this skill. It's read-only by design — propose, don't act. The user runs follow-up /memoryze / re-tag / delete operations themselves.
What you collect
Run these in order and aggregate:
# 1. Distribution health
graft stats
# 2. Usage signal across the lifetime of the graph
graft analytics
# 3. Last-week trend (compare with lifetime to spot degradation)
graft analytics --since 7d
# 4. Profile context
graft profile current
graft profile list
If the user said "audit profile X" or "tutti i profili", iterate with GRAFT_PROFILE=X per profile.
What you analyze
A. Hit-rate health
From analytics.cache.hit_rate:
- >= 0.30 — healthy mature graph.
- 0.10 to 0.30 — adequate, but may improve with better summaries.
- < 0.10 — either the graph is too young (< 50 queries) OR something's wrong: bad summaries, or the agent isn't searching before answering.
Use events.total to disambiguate: < 50 events ⇒ "too young to judge"; >= 50 ⇒ real problem.
B. Hoarding ratio
From analytics.insert_to_query_ratio:
- < 0.5 — healthy (more searches than saves; the graph is being USED).
- 0.5 - 1.5 — tolerable; the agent saves and searches in roughly equal measure.
- > 1.5 — the agent is hoarding without checking first. Tell the user: "we're saving more than searching — likely creating duplicates."
C. Champion vs. orphan nodes
From analytics.top_reused_nodes:
- The top 3-5 IDs by
hitsare champions. For each, show:graft get <id_hex>→ the full content- Suggestion: if
hits >= 5, consider promoting the content to a README / docs page — it's load-bearing knowledge.
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.
- 4d ago First seen · 138 lines · 125 tokens per session scan A 73bedf4159ca
memory-audit is a skill published in the GitHub repository AEndrix03/Graft (11 stars, last pushed 11d ago), licensed Apache-2.0. It adds 125 tokens to every session and 1,566 once invoked, about $0.0006 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 2026-08-30.
Other skills, from other repositories
Context Doctor
Identify and repair degradation in system prompt, external memory, and skills preventing you from following instructions or remembering information as well as you should.
memanto-companion
Inspect and manage the cross-session engineering memory that Memanto maintains for your Claude Code skills. Use when the user asks what Memanto remembers, wants to see their engineering profile, manually recall context for a skill, or store a decision. The automatic lifecycle hooks handle capture/injection on their…
memory-recall
Search and recall relevant memories from past sessions via memsearch. Use when the user's question could benefit from historical context, past decisions, debugging notes, previous conversations, or project knowledge -- especially questions like 'what did I decide about X', 'why did we do Y', or 'have I seen this…
mnemon
Persistent memory CLI for LLM agents. Store facts, recall past knowledge, link related memories, manage lifecycle.
mnemon
Persistent memory for MiniMax Code. Recall durable context, store important facts and decisions, and link related memories with the mnemon CLI.
prolong
Recover and use durable coding-session history from PRO-LONG's local append-only log. Use on long-running coding tasks, after context compaction or session resume, when reconstructing prior decisions or tool results, or before repeating work that may already have been attempted.