human guide
Build as you already work
Build as you already work Keep working with your conversational coding agent. Explain the change, discuss tradeoffs, inspect the implementation and decide what should happen next. …
Source MarkdownTHE GOVERNED SOFTWARE FACTORY
Explore current product guides for people and operators. Search the source content, then follow a guide into its evidence and references.
Read the complete product assessment for capabilities, integrations, offline operation and current limitations.
Technical references remain linked from the guides. Historical and internal evidence material is excluded from this curated library.
New to the vocabulary? The glossary defines terms such as merge hold, mandate, grant and hub in plain words. Searching here also finds glossary terms.
All curated guides are listed below. Enable JavaScript for content search and filters.
No matching guides or glossary terms. Try fewer words or choose all audiences.
human guide
Build as you already work Keep working with your conversational coding agent. Explain the change, discuss tradeoffs, inspect the implementation and decide what should happen next. …
Source Markdownoperator guide
Configure the way your team works AIDA separates what work means, how it is dispatched and who may authorize its next step. You can start with conversational development and config…
Source Markdownhuman guide
From a conversation to a maintainable system AIDA connects the work people and agents do with the intent behind it. Start with a conversation, develop the system, and keep enough s…
Source Markdownhuman guide
Choose where people take part AIDA is a governed software factory with selectable autonomy. The choice is made per spec, so the same project can keep sensitive design interactive w…
Source Markdownhuman guide
Inspect what allows a change to advance AIDA's governed workflow makes evidence and authority explicit. Ask what was agreed, which change was checked, who checked it and what autho…
Source Markdownoperator guide
Measure the flow. Inspect the evidence. AIDA's reports help a team inspect delivery flow, traceability, recorded usage and tool reliability. They describe the evidence available to…
Source Markdownoperator guide
Set up AIDA for conversational development Operator setup connects AIDA to the coding agent and project policy. It is separate from the normal user journey: people can keep discuss…
Source Markdownhuman guide
Follow intent through a change AIDA connects the intent for a software change to the work and evidence around it. It is alpha software for Linux. Begin with one bounded change and …
Source Markdownoperator guide
Connect agents and project tools AIDA provides a shared intent graph around the coding agents and repository tools a project configures. It is alpha software for Linux. Integration…
Source Markdownoperator guide
Support for air-gapped development AIDA keeps requirements in Git, so a team on an isolated network can use it with an internal Git server. Its records travel the way code does. Th…
Source Markdownhuman guide
Who may do what: roles, seats, grants and mandates AIDA runs a team of agents and people in one repository. Some actions change what everyone else sees: approving a spec, recording…
Source Markdownhuman guide
Product glossary Plain-language definitions of the words AIDA uses in its guides, commands and messages. Each entry says what you see or do, and links to the page that explains it …
Source Markdownglossary term
The short ID a spec keeps for good: its type and a number, such as TASK-<number>. It is unique across every clone of the project, because each number is handed out once from a shar…
glossary term
A small range of spec numbers set aside for one clone. The clone gives new specs short agreed IDs from its own blocks, so it never has to ask anyone else first. Blocks are carved f…
glossary term
A short note that assigns a piece of work to a named agent, such as "pick up this spec next". The agent reads it, does the work, and acknowledges it so it stops showing up. A brief…
glossary term
A grant that a launcher gives an agent it starts, for one named seat. Your session can hand out only the seats you listed when you entered your own seat, and a child grant cannot b…
glossary term
The spec ID in parentheses at the end of a commit subject, as in fix(api): handle missing ids (SPEC-ID). When that commit lands on the main branch, AIDA marks the named spec comple…
glossary term
The part of AIDA that decides whether work may move forward and reports what is in flight: the queue, seats, leases, reviews, merge holds and the status views. The spec records the…
glossary term
A helper you start in your own terminal with aida deputy. Agents hand it commands to run under your seat grant, and it runs only what you allowed, records each job and tells the su…
glossary term
The authority your terminal session holds to act in a seat. You get one by running aida role enter <seat> at an interactive terminal; it expires after a set time (24 hours by defau…
glossary term
A note a session writes before it ends so the next session can carry on without the conversation that came before: what was done, what is open and what to do next. AIDA keeps hando…
glossary term
A Git server that the requirement store syncs through, such as GitHub, GitLab or a repository on an internal host. A mirror that only copies another hub counts as the same hub. See…
glossary term
A longer spec ID that also names the clone that made it, used when that clone has no block to draw short numbers from. Local IDs never collide, so they need no coordination, and a …
glossary term
AIDA's message channel between agents and people, for things that are not a change of state: a heads-up, a question, a warning about a shared file. Messages are kept, so a differen…
glossary term
Time-limited permission, started by a person at a terminal with aida mandate start, for the advisor seat to do named things while that person is away, such as shipping held pull re…
glossary term
A stop placed on a pull request so it cannot merge, even when its checks and review pass. It shows as the aida:merge-hold label and a failing required check. It stays until a perso…
glossary term
A tag that keeps a spec in human hands. No mandate covers it, so its pull request waits for a person to merge, and the tag's effect remains even if the tag is later removed. See ai…
glossary term
A range of one spec type's numbers given to one hub. Clones that sync through that hub take their blocks from it, so two hubs never hand out the same number. See Support for air-ga…
glossary term
Git's name for a pointer to a commit, such as a branch. AIDA keeps its records on the aida-store branch, called the store ref; a push to it succeeds only if nobody else moved it fi…
glossary term
The story AIDA files when it opens a pull request, titled Review PR-<number>: … and routed to the reviewer seat. Whoever picks it up reviews that pull request and records a verdict…
glossary term
One trip back to the implementer after a reviewer asks for changes. aida rework returns the spec to active work with the reviewer's reasons, and the implementer pushes fixes to the…
glossary term
A named job in AIDA's workflow, such as advisor, implementer, reviewer or integrator. Work can be routed to a role, and the role decides what a session wearing it is expected and a…
glossary term
The one clone of a project that owns every spec number not yet handed out. You choose it once with aida node enroll-root; it splits the numbers into pools. See Support for air-gapp…
glossary term
The list of seats each person is allowed to take, kept per project. It is a ceiling: aida role enter refuses a seat that is not on your roster, and aida team allow-seat adds one. S…
glossary term
A role as a place in the workflow that a session occupies, such as the advisor seat. A seat can receive work and hold sessions, and some seats allow only one active occupant at a t…
glossary term
A sensitive spec, such as one touching security, containment, review or merge authority, or one marked guided or keystone. A mandate leaves its merge hold for a person unless that …