sase memory

Inspect and initialize SASE memory context

Reference for sase memory. Run sase memory --help for the exact output on this checkout.

usage: sase memory [-h] {history,init,list,log,read,show,web} ...

Inspect SASE memory context. With no subcommand, defaults to sase memory list.

Bare sase memory defaults to sase memory list.

Options:

Option	Description	Default
-h, --help	show this help message and exit	—

Subcommands:

Subcommand	Summary
sase memory history	Browse committed memory and instruction-file history
sase memory init	Create or refresh memory files, AGENTS.md, and provider shims
sase memory list	Show loaded, referenced, available, and missing memory files
sase memory log	Summarize or inspect auditable reference memory reads
sase memory read	Read and audit one or more memory selectors
sase memory show	Print one or more memory selectors without recording a read
sase memory web	List and inspect memory webs and their strands

sase memory history

usage: sase memory history [-h] [-a] [-A REV] [-d] [-f {json,pager,text}]
                           [-l N] [-m] [-p REF] [-s DATE]
                           [-S {all,home,project}]
                           [SELECTOR ...]

Browse the committed history of SASE memory notes, webs, strands, and agent instruction files (AGENTS.md plus its provider shims, project and home). With no selector, shows the cross-file changes feed; with selectors, shows each subject's timeline. Viewing history never records an audited read.

Options:

Option	Description	Default
-a, --all	Show hidden versions (moves, reflows) and regen-only changesets	false
-A REV, --at REV	Show one version: an ordinal (7 or v7), ~N (~1 is newest), a commit SHA prefix, blob:OID (a full blob OID or a unique prefix of at least 7 hex characters), now for the worktree file, or a YYYY-MM-DD date (latest version at or before it)	—
-d, --diff	Show the change instead of the body	false
-f FORMAT, --format FORMAT	Output format (default: pager on a TTY, else text). json emits the Rust wire unchanged; text prints timelines and bodies. (choices: json, pager, text)	—
-h, --help	show this help message and exit	—
-l N, --limit N	Keep at most N feed changesets or timeline versions	—
-m, --mark-reviewed	Mark the shown scopes reviewed through their newest changesets (feed mode only; an error with selectors)	false
-p REF, --project REF	Project to resolve memory from (default: infer from current directory)	—
-S SCOPE, --scope SCOPE	History scopes to query (default: all) (choices: all, home, project)	all
-s DATE, --since DATE	Only include changes at or after YYYY-MM-DD	—

Arguments:

Argument	Description
selectors	Memory or instruction selectors: a flat note name (tui.md), a repo-relative path (sase/memory/tui.md), a bare web name (glossary), a web:keyword strand reference (glossary:stitch), an instruction path (AGENTS.md, CLAUDE.md, subdir/AGENTS.md, ~/AGENTS.md), or a historical name (build_and_run.md). With no selector, shows the changes feed.

sase memory init

usage: sase memory init [-h] [-c] [-d] [-M] [-m MESSAGE] [-C]

Create or refresh SASE memory files, managed AGENTS.md, and provider instruction shims. sase init memory is a compatibility alias for this command.

Options:

Option	Description	Default
-c, --check	Report memory initialization drift without writing files	false
-d, --diff	Show full file diffs for planned memory changes	false
-M, --enable-project-memory	Create or update sase/sase.yml with is_sase_managed: true, enabling managed project memory (cannot be combined with --check)	false
-h, --help	show this help message and exit	—
-m MESSAGE, --message MESSAGE	Commit subject for folding eligible memory and generated-change source edits; a docs(memory): tag is added if omitted	—
-C, --no-commit	Skip the project git commit/push sequence	false

sase memory list

usage: sase memory list [-h]

Show the memory files visible from the current launch context, including loaded @ references, referenced-only plain memory paths, available files, and missing references.

Options:

Option	Description	Default
-h, --help	show this help message and exit	—

sase memory log

usage: sase memory log [-h] [--path MEMORY_PATH] [--agent AGENT_NAME]
                       [--id READ_ID] [--include KIND] [--json]

Summarize auditable reference memory reads recorded by sase memory read, or inspect matching read events with --path, --agent, or --id.

Options:

Option	Description	Default
--agent AGENT_NAME	Only include reads by the given agent	—
-h, --help	show this help message and exit	—
--id READ_ID	Show one memory read event by id or unambiguous id prefix	—
--include KIND	Include additional audit events: 'glossary' folds in the legacy sase glossary read audit log (choices: glossary)	—
--json	Emit deterministic machine-readable JSON	false
--path MEMORY_PATH	Only include reads for the given memory-relative path	—

sase memory read

usage: sase memory read [-h] [-d N] [-f {json,markdown,rich}] [-p REF]
                        -r REASON
                        selector [selector ...]

Resolve one or more memory selectors from project sase/memory/, falling back to ~/sase/memory/ when the project file is absent. A selector is a flat note name (generated_skills.md), a bare memory-web name (glossary, every strand), or a web:keyword strand reference (glossary:stitch). The whole batch is resolved before anything is printed or logged, so one unknown selector fails the entire request. Leading YAML frontmatter is stripped from notes, child notes are listed when present, and each read appends one attributable audit log row. Identical to sase memory show except that it requires a reason and records an audited read before printing.

Options:

Option	Description	Default
-d N, --depth N	Cap link/mention closure recursion depth (default: unlimited); -d 0 prints only the requested selectors and lists every link as a reference	—
-f FORMAT, --format FORMAT	Output format (default: markdown) (choices: json, markdown, rich)	markdown
-h, --help	show this help message and exit	—
-p REF, --project REF	Project to resolve memory from (default: infer from current directory)	—
-r REASON, --reason REASON	Non-empty reason for the audited memory read (required)	—

Arguments:

Argument	Description
selectors	One or more selectors: a flat note name (generated_skills.md), a bare memory-web name (glossary), or a web:keyword strand reference (glossary:stitch)

sase memory show

usage: sase memory show [-h] [-d N] [-f {json,markdown,rich}] [-p REF]
                        selector [selector ...]

Resolve one or more memory selectors exactly like sase memory read — project sase/memory/ first, then ~/sase/memory/ — strip leading YAML frontmatter from notes, and append the ## Children section when children exist. Unlike read, this command records no audit event, so agents consulting memory to do work must use sase memory read instead.

Options:

Option	Description	Default
-d N, --depth N	Cap link/mention closure recursion depth (default: unlimited); -d 0 prints only the requested selectors and lists every link as a reference	—
-f FORMAT, --format FORMAT	Output format (default: markdown) (choices: json, markdown, rich)	markdown
-h, --help	show this help message and exit	—
-p REF, --project REF	Project to resolve memory from (default: infer from current directory)	—

Arguments:

Argument	Description
selectors	One or more selectors: a flat note name (generated_skills.md), a bare memory-web name (glossary), or a web:keyword strand reference (glossary:stitch)

sase memory web

usage: sase memory web [-h] {list,show} ...

Inspect memory webs — keyed strand collections addressed with sase memory read <web>:<keyword>. Running sase memory web defaults to sase memory web list.

Bare sase memory web defaults to sase memory web list.

Options:

Option	Description	Default
-h, --help	show this help message and exit	—

Subcommands:

Subcommand	Summary
sase memory web list	List memory webs for a project
sase memory web show	Show one web's filterable strand index

sase memory web list

usage: sase memory web list [-h] [-f {json,names,table}] [-p REF]

List discovered memory webs: name, rendering (core/reference), scope, and strand count.

Options:

Option	Description	Default
-f FORMAT, --format FORMAT	Output format (default: table) (choices: json, names, table)	table
-h, --help	show this help message and exit	—
-p REF, --project REF	Project to resolve memory webs from (default: infer from current directory)	—

sase memory web show

usage: sase memory web show [-h] [-b] [-f {json,names,table}] [-p REF]
                            WEB [PATTERN]

Print one web's strand index: keyword, slug, aliases, mention-reference count, and summary, optionally filtered by PATTERN. This is the index; sase memory show <web>:<keyword> is the content read.

Options:

Option	Description	Default
-b, --bodies	Extend PATTERN matching into strand bodies	false
-f FORMAT, --format FORMAT	Output format (default: table) (choices: json, names, table)	table
-h, --help	show this help message and exit	—
-p REF, --project REF	Project to resolve memory webs from (default: infer from current directory)	—

Arguments:

Argument	Description
web	Memory web slug to inspect
pattern	Case-insensitive substring filter over keywords and aliases

Related: CLI command groups.

_[project]: A sase project is a named unit of work registered with SASE. _[workspace]: A sase workspace is a numbered clone of a project's primary repo, managed by the workspace store and tracked in that project's registry.json. _[agent]: A sase agent is an agent session or a single agent that does not belong to a session. _[prompt]: A prompt is the text SASE hands to a provider to start an agent turn. _[macro]: A macro is a reusable prompt triggered with #name in agent prompts. _[Patch]: A Patch is SASE's local unit of change. _[bead]: A bead is one tracked work item in SASE's git-native issue tracker. _[plan]: A plan is a persisted, approved artifact recording the intent behind agent work. _[TUI]: The TUI is SASE's terminal interface for running agents and reviewing results. _[epic]: An epic is an executable multi-phase plan run with sase bead work. _[gate]: A sase gate is a durable, command-backed request for one user decision, created when work must pause for the user. _[monitor]: A sase monitor is a monitor turn that runs one long command under a detached supervisor, so the command outlives the agent that started it. *[stitch]: A stitch is the lightweight ordered change record inside a Patch's STITCHES: section.