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 Yesterday-AI/skills --skill systemic-persistencegit clone --depth 1 https://github.com/Yesterday-AI/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/yesterday-ai/skills/systemic-persistence)<a href="https://agentmods.dev/skills/yesterday-ai/skills/systemic-persistence"><img src="https://agentmods.dev/badge/skills/yesterday-ai/skills/systemic-persistence.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 | $0.00051 | $0.00811 |
| Opus 5 | $0.00026 | $0.00405 |
| Sonnet 5 | $0.00010 | $0.00162 |
| Haiku 4.5 | $0.00005 | $0.00081 |
Grade A, and why
systemic-persistence 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 — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Systemic Persistence 💾🛡️
"Ein Fix, der einen Reboot nicht überlebt, ist kein Fix, sondern ein Workaround."
Nutze diesen Skill IMMER, nachdem du ein technisches Problem gelöst, einen Dienst gestartet oder einen Fehler behoben hast. Er zwingt dich, die Lösung nachhaltig und systemweit zu verankern.
Usage
Wende diese Checkliste unaufgefordert an, sobald du ein Problem für den User erfolgreich gelöst hast.
The Persistence Checklist ✅
Gehe diese Punkte systematisch durch. Wenn einer fehlt, setze ihn um oder informiere den User.
1. REBOOT_SURVIVAL 🔄
- Definition: Überlebt der Fix einen Systemneustart?
- Rule: Manuell gestartete Hintergrundprozesse (z.B. Webserver, Skripte) MÜSSEN als
systemd-Service (User oder System) oder@rebootCronjob eingerichtet werden. - Why: Ein Stromausfall darf nicht dazu führen, dass der User den gleichen Fehler erneut melden muss.
- Action: Schreibe eine
.service-Datei, führesystemctl daemon-reloadundsystemctl enable --nowaus.
2. LOG_ROTATION 📜
- Definition: Produziert die Lösung kontinuierlich Logs oder Datenmüll?
- Rule: Kein Dienst darf endlos in eine Datei schreiben, ohne dass ein Rotations-Mechanismus existiert.
- Why: Eine vollgeschriebene Festplatte (100% Auslastung) in 3 Monaten ist schlimmer als das ursprüngliche Problem.
- Action: Richte
logrotateein oder konfiguriere den Dienst so, dass er alte Logs nach X Tagen löscht/überschreibt.
3. EXPLICIT_MEMORY 🧠
- Definition: Ist die Lösung für die Zukunft dokumentiert?
- Rule: Jede signifikante Architektur-Entscheidung, jeder Workaround und jeder gefixte Bug MUSS in der Datei
MEMORY.mdoder der projektspezifischen Dokumentation festgehalten werden. - Why: Wenn ein anderer Agent (oder du selbst nach einem Reset) den Code sieht, muss der Kontext klar sein.
- Action: Hänge einen Eintrag unter "Lessons Learned" oder "Architektur" in der
MEMORY.mdan.
4. SCRIPTED_REPRODUCTION 🛠️
- Definition: Wenn das Problem auf einem anderen Rechner auftritt, lässt es sich mit einem Befehl lösen?
- Rule: Komplexe Kommandozeilen-Hacks (z.B. Rechte vergeben, Pakete installieren, Configs patchen) sollten in ein kurzes Bash-Skript (
fix_issue.shodersetup.sh) gegossen werden. - Why: Automatisierung schlägt manuelles Tippen.
- Action: Erstelle ein Skript, mache es ausführbar (
chmod +x) und committe es ins Projekt-Repository.
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 · 57 lines · 51 tokens per session scan A b52da4301201
systemic-persistence is a skill published in the GitHub repository Yesterday-AI/skills (2 stars, last pushed 4d ago), licensed MIT. It adds 51 tokens to every session and 811 once invoked, about $0.0003 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
design-agent-loop
Design a reusable agent loop as a graph before running it. Take a TASK and COMPLETION CRITERIA, draw the node graph (input, plan, act, verify, retry or escalate, done), specify context, tools, expected output, and evidence per node, label every transition trigger, set max retries, budget ceilings, and human approval…
gauntlet-loop
Beat a real-world quality bar with a Gauntlet Loop: take a GOAL and a REAL-WORLD EQUIVALENT with an inspectable reference pack, decompose into independent parts, assign specialist builders, and gate each part behind a separate fresh-context critic. Pass only if better than the reference; otherwise return the largest…
blind-spot-pass
Run a pre-build blind spot pass before the agent builds anything. Treat the user's prompt as the map and the real project as the territory: classify gaps across known knowns, known unknowns, unknown knowns, and unknown unknowns (with deps, edge cases, and blast radius under unknown unknowns); ask 5 to 10 high-leverage…
define-agent-goal
Turn a rough task into an agent Goal (outcome, verification with 3 to 5 success criteria, constraints, boundaries, iteration policy, stopping condition, and helper goals when parallel work is implied) so an agent can run with minimal babysitting. Plan-only, no execution in the same turn. Use for Goal, delegate, run…
merge-open-prs
Review, verify (Docker by default), and auto-merge open PRs when CI, reviews, and local verification pass. Batch up to 10 PRs per run. Use when the user wants to merge all open PRs, clear the PR queue, or run /merge-open-prs.
automation-roi-audit
Act as an AI operations consultant: interview how a business function's workflows actually run, label each step Human-only / AI-assisted / AI-owned, pick the highest-ROI automation tied to money, margin, or output, warn on AI theater and distraction traps, and recommend one narrow one-week test. Use for automation…