Discovery
Finding things, remembering things, and keeping a record of what you decided. Search and memory are the retrieval half; the Decision Ledger (decide / decisions) is the judgment half — a durable record of choices and the model positions that informed them.
search
Aliases: none
Purpose: Search tasks and docs.
Syntax:
search <query>
Flags: none
Examples:
search AV setup
See also: docs, memory, pull
Matches titles, tags, list and project names, and doc content.
decide
Aliases: none
Purpose: Record a decision — your choice, your reasoning, and the model positions behind it.
Syntax:
decide <choice> --why <rationale> [--confidence <0-1>] [--review <date>] [--doc <title>] [--question <text>]
Flags:
--why <rationale> why you chose this (required)
--confidence <0-1> how confident you are
--review <date> when to revisit the decision
--doc <title> the saved council/debate doc that informed it
--question <text> the question decided, if not obvious from the doc
Examples:
decide open with the story --why the demo can fail live --confidence 0.7
decide skip the second rehearsal --why diminishing returns --doc "Council: rehearsal plan"
See also: decisions, council
Shows a receipt and records only after you confirm with y. If you omit --confidence, the confirmation asks for it (y 0.7) — BEAM never invents a confidence number for you. Point --doc at a saved council or debate doc and the dissenting model positions are captured alongside your choice.
decisions
Aliases: none
Purpose: Browse recorded decisions.
Syntax:
decisions
decisions search <query>
decisions #N
Flags: none
Examples:
decisions
decisions search rehearsal
decisions #12
See also: decide, council
Each entry shows your choice, a confidence band, and how many models dissented. #N filters to decisions linked to a task. The Decisions panel on the right shows the same record as a timeline. For the raw transcripts behind a decision, use council history / council search.
memory
Aliases: none
Purpose: What BEAM remembers about you and your work.
Syntax:
memory
memory <query>
memory list [--category <name>]
memory add "content" --category <preference|pattern|decision|goal|relationship>
memory delete <number-or-text>
memory assumptions
memory test <ref> confirm|reject
memory reset <patterns|preferences|assumptions|category|all>
memory stats
Flags:
--category <name> filter or classify by category
Examples:
memory
memory add "prefers morning rehearsals" --category preference
memory test 3 reject
See also: portrait, search
Memories accrue two ways: you add them explicitly, or BEAM extracts them from your side of AI conversations. Each carries a confidence zone (● established, ◐ likely, ○ tentative, ◌ untested); assumptions lists the untested ones and test lets you confirm or reject them, which adjusts confidence. reset clears confidence without deleting content. Memories are encrypted like everything else.
shelf
Aliases: none
Purpose: Frictionless capture for half-formed ideas.
Syntax:
shelf
shelf "idea"
shelf this
shelf promote <n>
shelf drop <n>
shelf clear
Flags: none
Examples:
shelf "what if the demo runs on stage wifi — fallback?"
shelf this
shelf promote 2
See also: add, ask
shelf this shelves the last AI response. promote turns a shelved item into a real task when it's ready.
activity
Aliases: feed
Purpose: Recent activity in the workspace.
Syntax:
activity
Flags: none
Examples:
activity
See also: history
history
Aliases: none
Purpose: Recent command history.
Syntax:
history
Flags: none
Examples:
history
See also: activity
inspire
Aliases: none
Purpose: Print a random BEAM quote.
Syntax:
inspire
Flags: none
Examples:
inspire
See also: brief