MAIOS / Scheduled activities (AP) for autonomous systems

AP for project possibilities and opportunities

A scheduled activity that connects external developments with a project’s capabilities, identifying possibilities and opportunities with sources, conditions and useful follow-up.

The Radar observes changes around a project and within its work. It connects sources, needs, competences and resources to understand which possibilities become relevant: another way of working, a question to investigate or a contribution to offer others. It preserves the reasons and conditions for each possibility, follows earlier cases and reports differences that deserve attention.

Edition 0.1.0 · Available method with illustrative cases

How the Radar recognizes useful project possibilities

The project the Radar starts from

The Radar understands the project’s purpose, current work, established decisions and what its people can do. It reads available sources and possibilities retained from previous cycles. A new tool can be relevant, as can an emerging need, an accessible resource or a changed way of using an existing capability.

Sources and their connection to a concrete need

The AP searches and compares information through the means available to it. It distinguishes a source’s claims from what the project could do with them. For example, a cultural archive might encounter a text-recognition tool: the Radar clarifies which documents it supports, what work it could help with and which conditions remain to be understood.

Possibilities the project receives and contributions it can offer

The connection can work in both directions. An external development can open a new use within the project; one of its competences can become useful to another organization. The Radar preserves the possible recipient, reasons, sources, required resources and following question. Follow-up may be an investigation, a deferred possibility or an informative handoff to whoever will develop the work.

Possibilities over time and useful notifications

Possibilities remain in an archive that preserves reasons, conditions and changes. New information updates the same case when it concerns the same relationship. The AP can learn that a source was unsuitable or that a competence was missing from its reading and improve its method. Notifications follow the entrusted preferences; a cycle without useful differences can remain quiet.

Four cases of observing and following possibilities

Repair workshop: technical skills for a library of things

Illustrative case. A workshop can repair small appliances and observes a library planning to lend objects to its neighbourhood. The Radar connects the maintenance need with the workshop’s competences and retains a possible contribution: helping define checks and repair paths. People’s availability, organization and resources remain unknown. The note reaches the workshop’s responsible person; contacting the library requires a distinct mandate. The possibility concerns what the project could offer, without presuming interest or agreement.

Cultural archive: text recognition and handwritten documents

Illustrative case. An archive wants to make handwritten notebooks searchable and encounters an announcement for an automatic text-recognition tool, known as OCR. Reading the original guide, the Radar discovers that the examples concern printed text. It retains the difference and a useful question: which part of the collection could benefit and which capabilities are needed for manuscripts? Software already used by the archive may cover printed text. The possibility is reformulated; no installation or manuscript quality is attributed to the announcement.

Research group: a dataset changes a question about local space

Illustrative case. A group studies green-space use with daytime data. A new public source describes evening observations and makes another question about access routes thinkable. The Radar reaches the source description, period and conditions, then connects the possibility to the group’s competence. Coverage, compatibility and permitted use remain to be understood. The handoff is a research question with its sources; analysis, results and publication belong to subsequent scientific work.

Neighbourhood services: a resource reopens a deferred possibility

Illustrative case. A network had retained the idea of a shared service desk, deferred for lack of accessible space. A later source describes premises that might become available. The Radar updates the same case, preserves why it was deferred and identifies what changed. Actual access, timing, people and operation remain open conditions. It reports the new relationship according to entrusted preferences; another article about the same premises with no new information does not create another opportunity.

The Radar in your own LLM and recurring work

These instructions help the person choose a project, make sources accessible and entrust the Radar to their own LLM or autonomous system.

The project and the first Radar cycle

Name the project, its purpose, current work and available sources. Start with a single development or an earlier possibility if useful. Ask the Radar to clarify what could change, on which grounds and under which conditions. The observed field can evolve when a pertinent relationship emerges.

Instructions to copy into your own LLM

Prepare and copy the complete starter from this page or use START_CHAT_EN.txt. Add context before or during the conversation: the text already contains the method, guide and examples. The assistant uses what it knows and can reach, then asks for pertinent missing information. Copying works without downloading the package.

Sources, archive and competences in your environment

Make entrusted materials accessible. With persistent files, retain the AP folder and use ACTIVITY.md as its entry. The package includes four complete methods: Signal Intelligence qualifies sources and signals; Situated Possibility Formation understands connections and possibilities; Meta-Activity connects mandate and scheduling; Situated Activity Evolution improves usefulness and method through results. In a chat, retain and provide the continuity note again. Specify where the Radar will find possibilities and who could use them.

Radar scheduling and notifications

After a first cycle, choose an event or cadence, timezone, recipient and notifications. You need an activity-management program, the controller, that can execute the assignment in your environment. Read back the instructions, status and schedule actually saved. Scheduling depends on the means available in your system. Choose whether to remain quiet when nothing useful changes.

The result to read and the possibility’s follow-up

Check whether you can understand which relationship emerged, why it matters to the project and what is missing for continuation. A possibility can remain in the archive with its conditions or pass to whoever will perform distinct work. Sources, attribution and status stay recognizable; the guide’s cases are illustrative. Later cycles resume from the present and useful consequences.

Complete Radar instructions to copy into your own LLM

Prepare and copy the complete instructions into your own LLM. The context below is optional: add it now or let your assistant understand it through the conversation. The text includes the method, guide and examples; preparation takes place in the browser and any context you enter stays on your device.

Public competences included in the Radar

The package includes four complete public competences. The AP composes them with the receiving system’s competences; the complete instructions already transfer the operating method.

Origin and status of the Radar method

A MAIOS resource by Graziano Guiducci, composed from Signal Intelligence, Situated Possibility Formation, Meta-Activity and Situated Activity Evolution. This first edition transfers the Radar method with illustrative cases. Configuration, execution and results in the receiving system retain their own conditions and evidence.

Activity source and Apache-2.0 licence ↗

Scheduled activities (AP) for autonomous systems ↗