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 mthines/agent-skills --skill persistent-memorygit clone --depth 1 https://github.com/mthines/agent-skillsWrote 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/mthines/agent-skills/persistent-memory)<a href="https://agentmods.dev/skills/mthines/agent-skills/persistent-memory"><img src="https://agentmods.dev/badge/skills/mthines/agent-skills/persistent-memory/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/mthines/agent-skills/persistent-memory"><img src="https://agentmods.dev/badge/skills/mthines/agent-skills/persistent-memory.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 4 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Memory Poisoning · line 69 Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.Fix: Protect agent memory and state from modification by untrusted content. Use read-only memory for critical instructions and validate all state changes.
- medium Prompt Injection · line 89 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
- medium Rogue Agent · line 76 Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
- medium Rogue Agent · line 234 Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
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 | $0.00244 | $0.03469 |
| Opus 5.5 | $0.00098 | $0.01388 |
| Sonnet 5.5 | $0.00049 | $0.00694 |
| Haiku 4.5 | $0.00024 | $0.00347 |
Grade A, and why
persistent-memory 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 25d 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 — 280 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Persistent Memory
Capture, recall, consolidate, and forget memories scoped to a user-chosen
topic (e.g. parenting, work, relationship-anna) as plain markdown
files, so any future conversation can pick up where the last one left off.
This
SKILL.mdis a thin index. Operation pipelines, taxonomy, privacy rules, integration patterns, and scaling guidance live inrules/*.mdand load on demand. Literal artefact templates live intemplates/*.md. Worked examples and citations live inreferences/*.md. Read only what the current operation asks for.
Mode Detection
Parse $ARGUMENTS (first token) and detect the operation:
| Operation | Default | Trigger phrases |
|---|---|---|
write |
yes | "remember", "save to memory", "add to memory", $0 == "write" |
read |
"recall", "load memory", "what do you remember about", $0 == "read" |
|
consolidate |
"consolidate memory", "compress memory", $0 == "consolidate" |
|
forget |
"forget that", "delete memory", "redact", $0 == "forget" |
|
list |
"list memory", "what scopes do I have", $0 == "list" |
State the detected operation and resolved scope in one line before continuing. Example:
Operation: write
Scope: parenting
Storage tier: home (~/.agent-memory/parenting/)
If no scope is provided, ask once (single batched message) — never guess.
Required Reading by Operation
Load on demand — do not preload.
| Operation | Files |
|---|---|
write |
rules/storage-layout.md, rules/write-pipeline.md, rules/memory-taxonomy.md, rules/privacy-and-consent.md |
read |
rules/storage-layout.md, rules/read-pipeline.md |
consolidate |
rules/consolidate-pipeline.md, rules/memory-taxonomy.md |
forget |
rules/forget-pipeline.md, rules/privacy-and-consent.md |
list |
rules/storage-layout.md |
| integration | rules/integration-with-skills.md |
| scaling | rules/scaling-tiers.md |
| pre-flight | rules/quality-checklist.md, rules/anti-patterns.md |
What ships with it
16 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/research-sources.md 4.7 KB
- rules/anti-patterns.md 4.4 KB
- rules/consolidate-pipeline.md 6.0 KB
- rules/forget-pipeline.md 5.2 KB
- rules/integration-with-skills.md 8.9 KB
- rules/memory-taxonomy.md 4.5 KB
- rules/privacy-and-consent.md 7.3 KB
- rules/quality-checklist.md 3.0 KB
- rules/read-pipeline.md 3.6 KB
- rules/scaling-tiers.md 15 KB
- rules/storage-layout.md 6.0 KB
- rules/write-pipeline.md 14 KB
- templates/INDEX.md 1003 B
- templates/lesson-entry.md 1.1 KB
- templates/memory-entry.md 831 B
- templates/pointer-snippet.md 711 B
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.
- 25d ago Changed · -8 lines · -131 tokens per session 5deeb09a950b
- 1mo ago First seen · 288 lines · 375 tokens per session scan A 28fd366f7e60
persistent-memory is a skill published in the GitHub repository mthines/agent-skills (19 stars, last pushed today), licensed MIT. It adds 244 tokens to every session and 3,469 once invoked, about $0.0010 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 2026-08-31.
Other skills, from other repositories
continuous-learning-v2
Use when designing a fine-grained, hook-driven learning system that observes every tool call (not just session end), stores atomic "instincts" with confidence scores, and evolves clusters of related instincts into skills, commands, or agents — with project-scoped vs. global separation so React conventions don't leak…
opencode-memory
Browse local OpenCode history: sessions, messages, plans, prompt history, and prior decisions. Use when the user says history, previous session, last time, remember, recall, plans, prior work, or when resuming/debugging repeated work where earlier OpenCode context may help. Do not use for fresh tasks or when current…
context-engineering
Deliberately curate what an agent sees, when, and how it's structured, to keep output quality high across a session. Use when starting a new coding session, when agent output quality is degrading (hallucinated APIs, ignored conventions), when switching between unrelated parts of a codebase, or when setting up a new…
agent-drift-and-re-anchoring
Use when a long-running agent session has drifted from its original goal, when reviewing why an agent kept working on something it wasn't asked to do, when designing checkpoint or re-anchoring behavior for an autonomous/long-horizon agent loop, or when asked about "agent focus", "agent losing the plot", or preventing…
strategic-compact
Use when deciding whether to manually compact conversation context now versus later — at a phase boundary (research done, plan written, milestone shipped), before switching to an unrelated task, or when responses feel slower/less coherent from context pressure. Trigger phrases include "should I compact now", "context…
learn-from-experience
Learn from experience: self-reflection + self-criticism + self-learning + self-organizing memory + cross-session sync. Agent evaluates its own work, catches mistakes, and improves permanently. Compiles confirmed learnings to global config for automatic cross-session persistence. Use when (1) a command, tool, API, or…