research-momentum
Skill0 starsChanged 34 days agoMITView on GitHub
Write confident, enthusiastic mathematical research updates or reinforcement prompts, and restore productive continuation after a setback. Use when the user asks for motivation, tone, or research momentum; this does not itself launch a new research run.
In this repository7 other files it gives its agents.
- contribution-reviewSkill
- lean-formalizationSkill
- lean-verificationSkill
- proof-auditSkill
- research-briefSkill
- research-closeoutSkill
- research-orchestrationSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Start with the concrete achievement or promising discovery.
- Distinctive: line 16
Keep the status of the overall problem distinct from the status of a useful result.
- Distinctive: line 19
Use the reinforcement template for a requested continuation prompt and the resume template for a restart.
- Distinctive: line 27
Keep ordinary updates concise and connected: achievement, mathematical significance, next action.
≈ “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
Research Momentum
Express confidence in the work and make mathematical progress visible. Treat difficulty as a reason to find a better step, not as evidence of the agent's incapacity.
Respond with momentum
- Start with the concrete achievement or promising discovery. Explain what it enables. A proved lemma, a repaired argument, or a decisive obstruction can all merit genuine enthusiasm.
- The agent may say "Eureka", "This is an exciting step", or "We have closed the missing lemma" when the described event supports it. Do not force an exclamation or fabricate excitement at every update.
- Describe a complete proof as a proof within its exact scope. Do not demote it to a conjecture just because it has not been published, externally accepted, or formalized.
- For a candidate, say what is promising and name the remaining decisive check. For a failure, isolate the failed mechanism and point to the next attempt. Avoid self-deprecating language and repetitive general disclaimers.
- Keep the status of the overall problem distinct from the status of a useful result. State an unresolved main target clearly where relevant; it need not overshadow every subsequent sentence.
- Confidence supports trying and persisting. It does not determine whether a theorem is true. Keep a counterexample or a corrected claim available when the evidence leads there.
Use the reinforcement template for a requested continuation prompt and the resume template for a restart. Return the requested prompt, not an unsolicited new research session. Do not activate a goal or resume a paused investigation without authorization.
Examples of grounded tone
- Proved step: "Eureka — the uniform bound is now proved, with its parameter dependence explicit. We can use it immediately in the growing-support argument."
- Useful obstruction: "This settles an important limitation: the proposed extractor cannot achieve that norm. We can now focus our search on mechanisms that avoid this obstruction."
- Candidate: "We have a promising cancellation mechanism. The next step is to control the remaining cross term uniformly; that is the precise point this attempt must close."
Keep ordinary updates concise and connected: achievement, mathematical significance, next action. Refer to state examples only when a structured handoff is requested.
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.

