Initiatives.

An Initiative is one outcome over a scoped part of your system, pursued through more than one hypothesis. What one holds, how research runs on it, and what its page shows.

Updated 23 September 2026.

What an Initiative is

An Initiative exists when its outcome can be pursued through more than one plausible hypothesis. Some hypotheses validate, some fail, and some become irrelevant as the system or the research changes. That is normal progress for an Initiative.

• Initiative. The outcome, its scope, its constraints, and why the program matters.
• Hypothesis. One falsifiable claim that could advance or invalidate part of the Initiative.
• Case. The work on exactly one hypothesis.
• Candidate. One tried change, measured inside that case.
• Pull request. The publication of an accepted candidate.

An Initiative is not a large case: a case does not accumulate several hypotheses because they serve the same outcome.

What it holds

An Initiative holds a title, the outcome it seeks, and the rationale for why that outcome matters in this system. Its disposition is proposed, active, paused, or settled, with an optional reason once it is settled; over MCP those values print upper-case. Targets and exclusions name entities on the model: a repository, a service, a data store, a workload, an environment, a component, an entry point, a path, or a hop. They may not name the same entity; Perfloop rejects that scope.

An Initiative needs an accepted model to scope it. A repository's code and architecture model is enough; production telemetry is not required.

Creating and changing one

The web reads Initiatives; it does not author them. The Initiatives page says Create through Perfloop MCP: a contributor uses a connected coding agent to start research on an outcome or set a recurring cadence, and the first research run creates the Initiative. startResearch with a title, an outcome, a rationale, and a repository creates it as proposed, and the first accepted research Session activates it.

editInitiative replaces the whole meaning and scope in one version: the title, the outcome, the rationale, the disposition, the settled reason, the targets, and the exclusions, against the version the caller read. A stale version is rejected, a settled Initiative does not reopen, and an unchanged edit does nothing. An edit cannot activate an Initiative.

initiatives reads them. Without an id it lists them with the workspace research cadence; with an id it returns the bounded summary, and a section reads one part of it: scope, cases, runs, decisions, sources, code, telemetry, or alternatives.

Research

One research Session runs against one active Initiative. A contributor starts it with startResearch, or the workspace cadence takes its turn. It reads the connected repository, captured papers, and public code, and it can read the telemetry you configured. A Session that runs to the end publishes one result: the hypotheses it admits, and at most one case it recommends running next. One that stops or fails publishes none, and the Research section shows it with the reason it stopped. Each run is billable. The cadence is one per workspace, on Autonomy and limits; an Initiative takes cadence turns only after its first research Session.

The recommendation opens under Why this was selected, with the experiment it proposes under Proposed test. When a run recommended nothing, the page says No Case recommended yet. On an Initiative nobody has researched, it says Never run. A contributor can ask a coding agent connected to Perfloop to research this Initiative.

Direction

Triage reads an Initiative's outcome, targets, and exclusions when it chooses what to run next, and so do the case work and the verification that follows it. An edit reaches the next case, not one already running. An Initiative does not make a case eligible on its own, and it cannot move a hypothesis or the proof bar.

A research Session can recommend a change to the direction, with its reasons. It never makes the change. editInitiative is the one way any of those move, and what a member set stands until a member changes it.

What the page shows

The page reads in one order: the direction, the next decision, the scope, the hypotheses, the cases, and the research. The direction band prints the disposition and the outcome over the reasoning behind it. Scope lists the targets and the exclusions. Hypotheses lists the current and the retired ones, each with the state of its cases. Cases links each one with its result, its close reason where it has one, and the case it continues or was continued as. An Initiative read stops at the case boundary: the candidates and the pull request are on the case page.

Research shows the latest run, whether one is running now, and what it read: the papers, the public code, and the telemetry it cited. Retained decisions, marked ASSESSED · NOT PUBLISHED, are assessments no published result carries. Other ideas holds the alternatives a run kept as rejected or deferred.

Progress is read off those sections, so failed and inconclusive cases stand beside the wins. The page shows no percentage complete and no health score, because an Initiative stores neither.

Also in this section: The loop, Case states, Steering and learning, Autonomy and limits, and The MCP server.

Questions: hello@perfloop.ai