systems-leverage
Skill8 starsChanged 57 days agoMITView on GitHub
Find where a small push changes the whole system. Use when local fixes keep failing, "where should we intervene?", recurring organizational patterns (shifting the burden, escalation, tragedy of the commons), behavior driven by stocks/flows/delays, adoption stalling, or shared resources being depleted.
In this repository16 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- boundary-designSkill
- causal-auditSkill
- decision-bias-checkSkill
- estimationSkill
- evidence-synthesisSkill
- experience-and-transmissionSkill
- failure-forensicsSkill
- formal-correctnessSkill
- measurement-disciplineSkill
- narrative-sensemakingSkill
- normative-designSkill
- problem-reframingSkill
- representation-and-possibilitySkill
- structure-discoverySkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 44
If no shape above matches, use a built-in agent with the closest GENIUS: pattern from rules/agent-routes.json instead.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Systems Leverage
Problem shape: the system misbehaves as a whole while every part looks locally fine. Interventions at the parameter level keep failing because the structure — feedback loops, stocks, delays, governance rules — produces the behavior. The move: map the structure, rank the leverage points, intervene as high as is accessible.
Relevant geniuses
| Agent | Use when |
| meadows | many possible interventions, unclear where to focus; recurring archetype; confusing accumulation-and-delay behavior |
| beer | organization fails despite local fixes — audit the five viable-system functions; variety mismatch with the environment |
| kauffman | system too rigid or too chaotic — tune connectivity; explore the adjacent possible instead of leaping |
| deming | management reacts to noise; quality inspected in at the end instead of built into the process |
| ostrom | a shared resource (build pipeline, on-call, budget, commons) is being depleted — check the eight design principles, rules-in-use vs rules-on-paper |
| margulis | competing units would gain more by merging — symbiosis as an architectural option |
| rogers | a change/tool isn't being adopted — segment the adopters, diagnose the chasm, audit the innovation's attributes |
| schelling | macro pattern emerges from micro choices nobody intended — model the tipping points before moralizing the outcome |
Invocation
- Pick the best-fit agent above. If two or more fit, run
tools/genius-invoker.sh route "<problem>"and take the top ranked match. - Load it:
tools/genius-invoker.sh invoke <agent> "<problem>", then readagents/genius/<agent>.mdin full. - Apply the agent's
<workflow>step by step and answer in its<output-format>. Deliver the structural diagnosis (loop, stock, rule) before any parameter recommendation. - Typical chain: meadows maps and ranks leverage → beer audits the control
structure → rogers plans the adoption. Run via
tools/genius-invoker.sh compose meadows beer -- "<problem>". - If no shape above matches, use a built-in agent with the closest
GENIUS:pattern fromrules/agent-routes.jsoninstead.
Refuse when
- The requester wants a parameter tweak explicitly instead of the structural diagnosis the symptoms call for — name the mismatch.
- The "system" is a single component with a local bug (use failure-forensics).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.