Enum Commands 

pub enum Commands {
    Tui {
        once: bool,
        reset_state: bool,
        asciicast: Option<PathBuf>,
        data_dir: Option<PathBuf>,
        inline: bool,
        ui_height: u16,
        anchor: String,
        record_macro: Option<PathBuf>,
        play_macro: Option<PathBuf>,
        refresh: bool,
    },
    Index {

        full: bool,
        force_rebuild: bool,
        watch: bool,
        watch_once: Option<Vec<PathBuf>>,
        watch_interval: u64,
        semantic: bool,
        build_hnsw: bool,
        embedder: String,
        data_dir: Option<PathBuf>,
        json: bool,
        idempotency_key: Option<String>,
        progress_interval_ms: u64,
        no_progress_events: bool,
        robot_trace_ingest: bool,
    
},
    Completions {
        shell: Shell,
    },
    Man,
    RobotDocs {
        topic: RobotTopic,
    },
    Search {

        query: String,
        agent: Vec<String>,
        workspace: Vec<String>,
        limit: usize,
        offset: usize,
        json: bool,
        robot_meta: bool,
        fields: Option<Vec<String>>,
        max_content_length: Option<usize>,
        max_tokens: Option<usize>,
        request_id: Option<String>,
        cursor: Option<String>,
        display: Option<DisplayFormat>,
        data_dir: Option<PathBuf>,
        days: Option<u32>,
        today: bool,
        yesterday: bool,
        week: bool,
        since: Option<String>,
        until: Option<String>,
        aggregate: Option<Vec<String>>,
        explain: bool,
        dry_run: bool,
        timeout: Option<u64>,
        highlight: bool,
        source: Option<String>,
        sessions_from: Option<String>,
        mode: Option<SearchMode>,
        approximate: bool,
        model: Option<String>,
        rerank: bool,
        reranker: Option<String>,
        daemon: bool,
        no_daemon: bool,
        two_tier: bool,
        fast_only: bool,
        quality_only: bool,
        refresh: bool,
    
},
    Pack {

        query: String,
        agent: Vec<String>,
        workspace: Vec<String>,
        limit: usize,
        json: bool,
        fields: Option<Vec<String>>,
        max_tokens: usize,
        max_sessions: usize,
        max_evidence: usize,
        context_lines: usize,
        max_excerpt_chars: usize,
        request_id: Option<String>,
        display: Option<DisplayFormat>,
        data_dir: Option<PathBuf>,
        days: Option<u32>,
        today: bool,
        yesterday: bool,
        week: bool,
        since: Option<String>,
        until: Option<String>,
        source: Option<String>,
        sessions_from: Option<String>,
        mode: Option<SearchMode>,
        freshness_policy: String,
        freshness_window_seconds: i64,
        require_evidence: bool,
        explain_selection: bool,
        refresh: bool,
        timeout: Option<u64>,
    
},
    Stats {
        data_dir: Option<PathBuf>,
        json: bool,
        source: Option<String>,
        by_source: bool,
    },
    Diag {
        data_dir: Option<PathBuf>,
        json: bool,
        quarantine: bool,
        verbose: bool,
    },
    Status {
        data_dir: Option<PathBuf>,
        json: bool,
        robot_meta: bool,
        stale_threshold: u64,
    },
    Capabilities {
        json: bool,
    },
    Triage {
        data_dir: Option<PathBuf>,
        json: bool,
        stale_threshold: u64,
    },
    State {
        data_dir: Option<PathBuf>,
        json: bool,
        robot_meta: bool,
        stale_threshold: u64,
    },
    ApiVersion {
        json: bool,
    },
    Introspect {
        json: bool,
    },
    View {
        path: PathBuf,
        source: Option<String>,
        line: Option<usize>,
        context: usize,
        json: bool,
    },
    Health {
        data_dir: Option<PathBuf>,
        json: bool,
        robot_meta: bool,
        stale_threshold: u64,
    },
    Doctor {

        data_dir: Option<PathBuf>,
        json: bool,
        check: bool,
        fix: bool,
        repair: bool,
        cleanup: bool,
        archive_scan: bool,
        archive_normalize: bool,
        backups_list: bool,
        backups_verify: bool,
        backups_restore: bool,
        baseline_save: bool,
        baseline_diff: bool,
        baseline_update: bool,
        support_bundle: bool,
        support_bundle_verify: bool,
        archive_export: bool,
        archive_relocate: bool,
        archive_export_verify: bool,
        archive_target: Option<PathBuf>,
        backup_id: Option<String>,
        baseline_id: Option<String>,
        baseline_path: Option<PathBuf>,
        support_bundle_path: Option<PathBuf>,
        include_sensitive_attachments: bool,
        sensitive_attachment: Vec<PathBuf>,
        sensitive_attachment_max_bytes: u64,
        dry_run: bool,
        yes: bool,
        plan_fingerprint: Option<String>,
        verbose: bool,
        force_rebuild: bool,
        allow_repeated_repair: bool,
        ls: bool,
        undo: Option<String>,
        robot_triage: bool,
        diff: Option<String>,
        gc_before: Option<String>,
        watch: bool,
        watch_interval_ms: u64,
        watch_iterations: u64,
        explain: Option<String>,
        emit_capabilities: bool,
    
},
    Context {
        path: PathBuf,
        source: Option<String>,
        data_dir: Option<PathBuf>,
        limit: usize,
        json: bool,
    },
    Sessions {
        workspace: Option<PathBuf>,
        current: bool,
        limit: Option<usize>,
        json: bool,
        data_dir: Option<PathBuf>,
    },
    Resume {
        path: PathBuf,
        agent: Option<String>,
        exec: bool,
        shell: bool,
        json: bool,
    },
    Upgrade {
        check: bool,
        force: bool,
        yes: bool,
        json: bool,
    },
    Export {
        path: PathBuf,
        source: Option<String>,
        format: ConvExportFormat,
        output: Option<PathBuf>,
        clipboard: bool,
        include_tools: bool,
        include_skills: bool,
    },
    ExportHtml {

        session: PathBuf,
        source: Option<String>,
        output_dir: Option<PathBuf>,
        filename: Option<String>,
        encrypt: bool,
        password_stdin: bool,
        include_tools: bool,
        show_timestamps: bool,
        no_cdns: bool,
        include_skills: bool,
        theme: String,
        dry_run: bool,
        explain: bool,
        open: bool,
        json: bool,
    
},
    Expand {
        path: PathBuf,
        source: Option<String>,
        line: usize,
        context: usize,
        json: bool,
    },
    Timeline {
        since: Option<String>,
        until: Option<String>,
        today: bool,
        agent: Vec<String>,
        data_dir: Option<PathBuf>,
        json: bool,
        group_by: TimelineGrouping,
        source: Option<String>,
    },
    Pages {

        export_only: Option<PathBuf>,
        verify: Option<PathBuf>,
        agents: Option<Vec<String>>,
        workspaces: Option<Vec<String>>,
        since: Option<String>,
        until: Option<String>,
        path_mode: PathMode,
        target: Option<PagesDeployTarget>,
        project: Option<String>,
        branch: Option<String>,
        account_id: Option<String>,
        api_token: Option<String>,
        dry_run: bool,
        scan_secrets: bool,
        fail_on_secrets: bool,
        secrets_allow: Vec<String>,
        secrets_deny: Vec<String>,
        json: bool,
        verbose: bool,
        no_encryption: bool,
        i_understand_unencrypted_risks: bool,
        preview: Option<PathBuf>,
        port: u16,
        no_open: bool,
        config: Option<String>,
        validate_config: bool,
        example_config: bool,
    
},
    Mirror(MirrorCommand),
    Sources(SourcesCommand),
    Models(ModelsCommand),
    Swarm(SwarmCommand),
    Import(ImportCommand),
    Analytics(AnalyticsCommand),
    Daemon {
        socket: Option<PathBuf>,
        idle_timeout: Option<u64>,
        max_connections: Option<usize>,
        data_dir: Option<PathBuf>,
    },
}

Variants

Tui

Launch interactive TUI

Fields

once: bool

Render once and exit (headless-friendly)

reset_state: bool

Delete persisted UI state (tui_state.json) before launch

asciicast: Option<PathBuf>

Record terminal output to an asciicast v2 file; in non-interactive headless –once mode cass writes a labeled sentinel cast because no live TUI session is launched.

data_dir: Option<PathBuf>

Override data dir (matches index –data-dir)

inline: bool

Run in inline mode (UI anchored within terminal, scrollback preserved)

ui_height: u16

Height of the inline UI in rows (default: 12, ignored without –inline)

anchor: String

Anchor the inline UI to top or bottom of the terminal (default: bottom)

record_macro: Option<PathBuf>

Record input events to a macro file for replay/debugging

play_macro: Option<PathBuf>

Play back a previously recorded macro file

refresh: bool

Run an incremental cass index pass before launching the TUI so new conversations created since the last index are searchable. No-op when the index is already current; indexing errors are logged and the TUI opens on the existing index (non-fatal).

Index

Run indexer

Fields

full: bool

Perform full rebuild

force_rebuild: bool

Force Tantivy index rebuild even if schema matches

watch: bool

Watch for changes and reindex automatically

watch_once: Option<Vec<PathBuf>>

Trigger a single watch cycle for specific paths (comma-separated or repeated)

watch_interval: u64

Minimum seconds between watch scan cycles (default: 30). Prevents high CPU usage from tight-loop scanning when filesystem events arrive continuously.

semantic: bool

Build semantic vector index after text indexing

build_hnsw: bool

Build HNSW index for approximate nearest neighbor search (requires –semantic). Enables O(log n) search with --approximate flag at query time.

embedder: String

Embedder to use for semantic indexing (hash, fastembed)

data_dir: Option<PathBuf>

Override data dir (index + db). Defaults to platform data dir.

json: bool

Output as JSON (for automation)

idempotency_key: Option<String>

Idempotency key for safe retries. If the same key is used with identical parameters, the cached result is returned. Keys expire after 24 hours.

progress_interval_ms: u64

Interval (ms) between NDJSON progress events emitted on stderr in –json/–robot mode. Clamped to [250, 60000]. Default 2000. Set –no-progress-events to disable.

no_progress_events: bool

Suppress NDJSON progress events on stderr in –json/–robot mode. Also honored via CASS_INDEX_NO_PROGRESS_EVENTS=1 env var.

robot_trace_ingest: bool

Emit per-ingest-batch NDJSON timing and lookup counters on stderr for perf bisection.

Completions

Generate shell completions to stdout

Fields

shell: Shell

Man

Generate man page to stdout

RobotDocs

Machine-focused docs for automation agents

Fields

topic: RobotTopic

Topic to print

Search

Run a one-off search and print results to stdout

Fields

query: String

The query string

agent: Vec<String>

Filter by agent slug (can be specified multiple times)

workspace: Vec<String>

Filter by workspace path (can be specified multiple times)

limit: usize

Max results. 0 = “no limit” but is auto-capped to a RAM-proportional ceiling (1/16 of MemAvailable, clamped to [256 MiB, 16 GiB] of result-heap) so a single query can’t tie up the whole machine. Override with CASS_SEARCH_NO_LIMIT_CAP= or CASS_SEARCH_NO_LIMIT_BYTES=.

offset: usize

Offset for pagination (start at Nth result)

json: bool

Output as JSON (–robot also works). Equivalent to –robot-format json

robot_meta: bool

Include extended metadata in robot output (elapsed_ms, wildcard_fallback, cache_stats)

fields: Option<Vec<String>>

Select specific fields in JSON output (comma-separated). Use ‘minimal’ for source_path,line_number,agent or ‘summary’ for source_path,line_number,agent,title,score. Example: –fields source_path,line_number

max_content_length: Option<usize>

Truncate content/snippet fields to max N characters (UTF-8 safe, adds ‘…’ and _truncated indicator)

max_tokens: Option<usize>

Soft token budget for robot output (approx; 4 chars ≈ 1 token). Adjusts truncation.

request_id: Option<String>

Request ID to echo in robot _meta for correlation

cursor: Option<String>

Cursor for pagination (base64-encoded offset/limit payload from previous result)

display: Option<DisplayFormat>

Human-readable display format: table (aligned columns), lines (one-liner), markdown

data_dir: Option<PathBuf>

Override data dir

days: Option<u32>

Filter to last N days

today: bool

Filter to today only

yesterday: bool

Filter to yesterday only

week: bool

Filter to last 7 days

since: Option<String>

Filter to entries since ISO date (YYYY-MM-DD or YYYY-MM-DDTHH:MM:SS), keyword (today, yesterday, now), or relative offset (-7d, -24h, -30m, -1w). allow_hyphen_values lets the dash-prefixed forms pass clap without requiring the equals syntax (reality-check bead hr0z4).

until: Option<String>

Filter to entries until ISO date / keyword / relative offset.

aggregate: Option<Vec<String>>

Server-side aggregation by field(s). Comma-separated: agent,workspace,date,match_type Returns buckets with counts instead of full results. Use with –limit to get both.

explain: bool

Include query explanation in output (shows parsed query, index strategy, cost estimate)

dry_run: bool

Validate and analyze query without executing (returns explanation, estimated cost, warnings)

timeout: Option<u64>

Timeout in milliseconds. Returns partial results and error if exceeded.

highlight: bool

Highlight matching terms in output (uses bold markers in text, in HTML)

source: Option<String>

Filter by source: ‘local’, ‘remote’, ‘all’, or a specific source hostname

sessions_from: Option<String>

Filter to sessions from file (one path per line). Use ‘-’ for stdin. Enables chained searches: cass search "query1" --robot-format sessions | cass search "query2" --sessions-from -

mode: Option<SearchMode>

Search mode: hybrid-preferred (default), lexical, or semantic

approximate: bool

Use approximate nearest neighbor (ANN) search with HNSW for faster semantic/hybrid queries. Trades slight accuracy loss for O(log n) search complexity instead of O(n). Only affects semantic and hybrid modes; ignored for lexical search. Requires an HNSW index built with cass index --semantic --approximate.

model: Option<String>

Embedding model to use for semantic search. Available models depend on what’s been downloaded. Use cass models --list to see available options.

rerank: bool

Enable reranking of search results for improved relevance. Requires a reranker model to be available.

reranker: Option<String>

Reranker model to use (requires –rerank). Use cass models --list to see available options.

daemon: bool

Use daemon for warm model inference (faster repeated queries). If daemon is unavailable, falls back to direct inference.

no_daemon: bool

Disable daemon usage even if available (force direct inference).

two_tier: bool

Enable two-tier progressive search: fast results immediately, refined via daemon. Returns initial results from fast embedder (~1ms), then refines with quality embedder via daemon (~130ms). Best of both worlds for interactive search.

fast_only: bool

Fast-only search: use lightweight embedder for instant results, no refinement. Ideal for real-time search-as-you-type scenarios where latency is critical.

quality_only: bool

Quality-only search: wait for full transformer model results. Higher latency (~130ms) but most accurate semantic matching. Requires daemon to be available; falls back to fast if unavailable.

refresh: bool

Run an incremental cass index pass before the search so new conversations created since the last index are matched. No-op when the index is already current; indexing errors are logged and the search runs against the existing index (non-fatal).

Pack

Build a deterministic answer pack for agent handoffs

Fields

query: String

The query string

agent: Vec<String>

Filter by agent slug (can be specified multiple times)

workspace: Vec<String>

Filter by workspace path (can be specified multiple times)

limit: usize

Max candidate hits to fetch before answer-pack planning. 0 uses the planner default.

json: bool

Output as JSON (–robot also works). Equivalent to –robot-format json.

fields: Option<Vec<String>>

Select answer-pack top-level fields. Presets: minimal, summary, all.

max_tokens: usize

Soft pack token budget.

max_sessions: usize

Maximum distinct sessions represented in the pack.

max_evidence: usize

Maximum evidence items selected into the pack.

context_lines: usize

Context lines requested around evidence hits.

max_excerpt_chars: usize

Maximum excerpt characters per evidence item.

request_id: Option<String>

Request ID to echo in pack metadata.

display: Option<DisplayFormat>

Human-readable display format. Only markdown is supported for pack.

data_dir: Option<PathBuf>

Override data dir

days: Option<u32>

Filter to last N days

today: bool

Filter to today only

yesterday: bool

Filter to yesterday only

week: bool

Filter to last 7 days

since: Option<String>

Filter to entries since ISO date (YYYY-MM-DD or YYYY-MM-DDTHH:MM:SS).

until: Option<String>

Filter to entries until ISO date / keyword / relative offset.

source: Option<String>

Filter by source: ‘local’, ‘remote’, ‘all’, or a specific source hostname.

sessions_from: Option<String>

Filter to sessions from file (one path per line). Use ‘-’ for stdin.

mode: Option<SearchMode>

Search mode: hybrid-preferred (default), lexical, or semantic.

freshness_policy: String

Freshness policy for evidence selection.

freshness_window_seconds: i64

Evidence freshness window in seconds.

require_evidence: bool

Return an error instead of an empty successful pack.

explain_selection: bool

Include selection score component details in evidence.

refresh: bool

Run an incremental cass index pass before packing.

timeout: Option<u64>

Timeout in milliseconds.

Stats

Show statistics about indexed data

Fields

data_dir: Option<PathBuf>

Override data dir

json: bool

Output as JSON (--robot also works)

source: Option<String>

Filter by source: ‘local’, ‘remote’, ‘all’, or a specific source hostname

by_source: bool

Show breakdown by source

Diag

Output diagnostic information for troubleshooting

Fields

data_dir: Option<PathBuf>

Override data dir

json: bool

Output as JSON (--robot also works)

quarantine: bool

Include quarantine and retained-asset inspection details

verbose: bool

Include verbose information (file sizes, timestamps)

Status

Quick health check for agents: index freshness, db stats, recommended action

Fields

data_dir: Option<PathBuf>

Override data dir

json: bool

Output as JSON (--robot also works)

robot_meta: bool

Include _meta block (elapsed, freshness, data_dir/db_path)

stale_threshold: u64

Staleness threshold in seconds (default: 1800 = 30 minutes)

Capabilities

First-stop agent self-description: workflows, mistake recoveries, commands, flags, env vars, exit codes, and limits

Fields

json: bool

Output as JSON (--robot also works)

Triage

One-shot agent triage: readiness, next command, workflows, docs, and schemas

Fields

data_dir: Option<PathBuf>

Override data dir

json: bool

Output as JSON (--robot also works)

stale_threshold: u64

Staleness threshold in seconds (default: 300)

State

Quick state/health check (alias of status)

Fields

data_dir: Option<PathBuf>

Override data dir

json: bool

Output as JSON (--robot also works)

robot_meta: bool

Include _meta block (elapsed, freshness, data_dir/db_path)

stale_threshold: u64

Staleness threshold in seconds (default: 1800 = 30 minutes)

ApiVersion

Show API + contract version info

Fields

json: bool

Output as JSON (--robot also works)

Introspect

Full API schema introspection - commands, arguments, and response schemas

Fields

json: bool

Output as JSON (--robot also works)

View

View a source file at a specific line (follow up on search results)

Fields

path: PathBuf

Path to the source file

source: Option<String>

Exact source_id from search output (e.g. ‘local’, ‘work-laptop’)

line: Option<usize>

Line number to show (1-indexed)

context: usize

Number of context lines before/after

json: bool

Output as JSON (--robot also works)

Health

Minimal health check (<50ms). Exit 0=healthy, 1=unhealthy. For agent pre-flight checks.

Fields

data_dir: Option<PathBuf>

Override data dir

json: bool

Output as JSON ({"healthy": bool, "latency_ms": N})

robot_meta: bool

Include _meta block (elapsed, freshness, data_dir/db_path)

stale_threshold: u64

Staleness threshold in seconds (default: 300)

Doctor

Diagnose cass installation issues. Legacy cass doctor --json maps to the read-only check surface. Legacy --fix maps to safe-auto-run and may only apply contract-declared safe repairs.

Fields

data_dir: Option<PathBuf>

Override data dir

json: bool

Output as JSON (--robot also works)

check: bool

Run the bounded read-only doctor truth surface (cass doctor check --json)

fix: bool

Legacy safe auto-run: apply only contract-declared safe repairs, preserving archive/source evidence.

repair: bool

Hidden normalized form for cass doctor repair ....

cleanup: bool

Hidden normalized form for cass doctor cleanup ....

archive_scan: bool

Hidden normalized form for cass doctor archive-scan ....

archive_normalize: bool

Hidden normalized form for cass doctor archive-normalize ....

backups_list: bool

Hidden normalized form for cass doctor backups list ....

backups_verify: bool

Hidden normalized form for cass doctor backups verify ....

backups_restore: bool

Hidden normalized form for cass doctor backups restore ....

baseline_save: bool

Hidden normalized form for cass doctor baseline save ....

baseline_diff: bool

Hidden normalized form for cass doctor baseline diff ....

baseline_update: bool

Hidden normalized form for cass doctor baseline update ....

support_bundle: bool

Hidden normalized form for cass doctor support-bundle ....

support_bundle_verify: bool

Hidden normalized form for cass doctor support-bundle verify ....

archive_export: bool

Hidden normalized form for cass doctor archive export ....

archive_relocate: bool

Hidden normalized form for cass doctor archive relocate ....

archive_export_verify: bool

Hidden normalized form for cass doctor archive export verify ....

archive_target: Option<PathBuf>

Target directory for cass doctor archive export|relocate.

backup_id: Option<String>

Backup id from cass doctor backups list --json.

baseline_id: Option<String>

Baseline id from cass doctor baseline save|diff <id>.

baseline_path: Option<PathBuf>

Explicit baseline file path for save/diff/update.

support_bundle_path: Option<PathBuf>

Explicit support bundle directory or manifest path for verify.

include_sensitive_attachments: bool

Include explicitly provided sensitive attachments in the support bundle.

sensitive_attachment: Vec<PathBuf>

Sensitive attachment path for cass doctor support-bundle --include-sensitive-attachments.

sensitive_attachment_max_bytes: u64

Maximum total sensitive attachment bytes copied into a support bundle.

dry_run: bool

Preview a fingerprinted repair plan without mutating any cass files.

yes: bool

Confirm a previously inspected fingerprinted repair plan.

plan_fingerprint: Option<String>

Plan fingerprint from cass doctor repair --dry-run --json.

verbose: bool

Run all checks verbosely (show passed checks too)

force_rebuild: bool

Request a derived rebuild; never bypasses archive coverage gates or plan fingerprints.

allow_repeated_repair: bool

Permit a mutating repair even when a previous failure marker exists

ls: bool

List per-run artifact directories under <data_dir>/doctor/runs/. Read-only. Output is JSON when combined with --json/--robot. Returns the run-id, started/ended timestamps, mode, exit_code, action_count and status (completed|incomplete|unknown) per run, newest-first.

undo: Option<String>

Undo the named run by id (or latest). Walks <run-dir>/actions.jsonl in reverse, verifies hashes, and restores byte-identically from the per-run backups. Refuses on any post-mutation tampering. The undo itself produces a fresh run-id.

robot_triage: bool

Mega-command for agents: returns one envelope with summary, findings, actions_planned, recommended_command, capabilities_url. Read-only; composes the existing read-only doctor check with planning hints.

diff: Option<String>

Show what --fix would change vs the current state. Read-only. Optional <REF> compares against a prior run-id instead.

gc_before: Option<String>

Quarantine doctor runs older than <ISO8601>. Renames into <data_dir>/doctor/quarantine/runs/; never deletes. Requires both --gc-before AND --yes (per the destructive-action gate).

watch: bool

World-class-doctor pass-6: long-running monitor mode. Periodically re-reads <data_dir>/doctor/runs/ and emits one JSONL event per observed change to stdout. Stops on Ctrl+C or after --watch-iterations.

watch_interval_ms: u64

World-class-doctor pass-6: poll interval for --watch. Default 5000 ms. Bounded to [500, 60000].

watch_iterations: u64

World-class-doctor pass-6: stop after this many polls. 0 = run forever (until SIGINT). Useful in tests + bounded automation.

explain: Option<String>

World-class-doctor pass-8: explain a prior run by id (or latest). Returns a single envelope with the run’s manifest, every action recorded, and a flattened band timeline. Read-only.

emit_capabilities: bool

World-class-doctor pass-8: emit the doctor’s extended capability surface — detectors[], fixers[], exit_codes[], data_paths[], env_vars[] — in one envelope. Read-only.

Context

Find related sessions for a given source path

Fields

path: PathBuf

Path to the source session file

source: Option<String>

Exact source_id from search output (e.g. ‘local’, ‘work-laptop’)

data_dir: Option<PathBuf>

Override data dir

limit: usize

Maximum results per relation type (default: 5)

json: bool

Output as JSON (--robot also works)

Sessions

List recent sessions, with optional workspace/current-session filtering

Fields

workspace: Option<PathBuf>

Filter to sessions for this workspace/project directory

current: bool

Resolve the current workspace automatically and return the most recent match

limit: Option<usize>

Maximum sessions to return (defaults: 10, or 1 with –current)

json: bool

Output as JSON (for automation)

data_dir: Option<PathBuf>

Override data dir

Resume

Resolve a session path into a ready-to-run resume command for its native harness (Claude Code, Codex, OpenCode, pi_agent, Gemini).

By default, the resolved command is printed to stdout, one argv token per line, so the caller can wrap it however they like:

eval “$(cass resume /path/to/session.jsonl –shell)” $(cass resume /path/to/session.jsonl –shell) # direct exec

Use --exec to have cass exec the command directly, replacing the current process.

Fields

path: PathBuf

Session file path (as printed by cass search or cass sessions)

agent: Option<String>

Override the detected harness (see --help for accepted values).

exec: bool

Replace the current process with the resolved resume command. Mutually exclusive with --shell and --json.

shell: bool

Emit a single shell-escaped command line on stdout (suitable for eval "$(cass resume ...)"). Mutually exclusive with --json.

json: bool

Output as JSON (for automation)

Upgrade

Check for a newer cass release and (optionally) install it.

Without flags, prompts y/N before running the same checksum- verified installer the TUI uses. Useful in headless environments where cass tui would never fire its own update prompt.

Examples:

cass upgrade # interactive prompt cass upgrade –check # print versions, exit non-zero if outdated cass upgrade –yes # install without asking cass upgrade –force –check # skip the 1-hour cadence

Fields

check: bool

Print current vs latest version and exit. No install. Exits 0 when up to date, 1 when an update is available.

force: bool

Bypass the 1-hour update-check cadence and re-fetch the GitHub release API immediately. Combine with --check to refresh status without installing.

yes: bool

Skip the interactive prompt and run the installer immediately. Suitable for scripts; mutually exclusive with --check.

json: bool

Output as JSON (for automation). Combines naturally with --check; ignored when an install actually runs because the installer execs over the current process.

Export

Export a conversation to markdown or other formats

Fields

path: PathBuf

Path to session file

source: Option<String>

Exact source_id from search output (e.g. ‘local’, ‘work-laptop’)

format: ConvExportFormat

Output format

output: Option<PathBuf>

Output file (stdout if not specified)

clipboard: bool

Copy the formatted export to the system clipboard instead of printing to stdout. Falls back to stdout with a stderr warning when no clipboard tool is available (e.g. headless / SSH).

include_tools: bool

Include tool use details in export

include_skills: bool

Include skill content in export (default: stripped for privacy)

ExportHtml

Export session as beautiful, self-contained HTML (with optional encryption)

Fields

session: PathBuf

Path to session file

source: Option<String>

Exact source_id from search output (e.g. ‘local’, ‘work-laptop’)

output_dir: Option<PathBuf>

Output directory (default: current directory)

filename: Option<String>

Custom filename (default: auto-generated from session metadata)

encrypt: bool

Enable password encryption (Web Crypto compatible)

password_stdin: bool

Read password from stdin (secure, no echo)

include_tools: bool

Include tool calls in export (default: true)

show_timestamps: bool

Show message timestamps

no_cdns: bool

Disable CDN references (fully offline, larger file)

include_skills: bool

Include skill content in export (default: stripped for privacy). Skills injected by Claude Code/Codex contain proprietary SKILL.md content that should not appear in shared/published exports.

theme: String

Default theme (dark or light)

dry_run: bool

Validate without writing file

explain: bool

Show export plan without executing

open: bool

Open file in browser after export

json: bool

JSON output (for automation)

Expand

Show messages around a specific line in a session file

Fields

path: PathBuf

Path to session file

source: Option<String>

Exact source_id from search output (e.g. ‘local’, ‘work-laptop’)

line: usize

Line number to show context around

context: usize

Number of messages before/after (default: 3)

json: bool

Output as JSON (--robot also works)

Timeline

Show activity timeline for a time range

Fields

since: Option<String>

Start time (ISO date, ‘today’, ‘yesterday’, ‘Nd’ for N days ago, or relative -7d/-24h/-30m/-1w). allow_hyphen_values lets dash-prefixed offsets pass clap (reality-check bead hr0z4).

until: Option<String>

End time (ISO date, keyword, or relative offset).

today: bool

Show today only

agent: Vec<String>

Filter by agent (can be repeated)

data_dir: Option<PathBuf>

Override data dir

json: bool

Output as JSON (–robot also works). Equivalent to –robot-format json

group_by: TimelineGrouping

Group by: hour, day, or none

source: Option<String>

Filter by source: ‘local’, ‘remote’, ‘all’, or a specific source hostname

Pages

Export encrypted searchable archive for static hosting (P4.x)

Fields

export_only: Option<PathBuf>

Export only (skip wizard and encryption) to specified directory

verify: Option<PathBuf>

Verify an existing export bundle (for CI/CD)

agents: Option<Vec<String>>

Filter by agent (comma-separated)

workspaces: Option<Vec<String>>

Filter by workspace (comma-separated)

since: Option<String>

Filter entries since ISO date, keyword, or relative offset (-7d, -1w, etc.). allow_hyphen_values lets dash-prefixed values pass clap (reality-check bead hr0z4).

until: Option<String>

Filter entries until ISO date, keyword, or relative offset.

path_mode: PathMode

Path mode: relative (default), basename, full, hash

target: Option<PagesDeployTarget>

Deployment target: local, github, cloudflare

project: Option<String>

Cloudflare project name (also used for GitHub repo name)

branch: Option<String>

Cloudflare production branch (default: main)

account_id: Option<String>

Cloudflare account ID (or CLOUDFLARE_ACCOUNT_ID env)

api_token: Option<String>

Cloudflare API token (or CLOUDFLARE_API_TOKEN env)

dry_run: bool

Dry run (don’t write files)

scan_secrets: bool

Scan for secrets and exit (no export)

fail_on_secrets: bool

Fail with non-zero exit if secrets are detected (for CI)

secrets_allow: Vec<String>

Allowlist regex patterns to suppress findings (repeatable or comma-separated)

secrets_deny: Vec<String>

Denylist regex patterns to force findings (repeatable or comma-separated)

json: bool

Output results as JSON (for verify and secret scan)

verbose: bool

Verbose output (show detailed check results)

no_encryption: bool

Export without encryption (DANGEROUS - all content publicly readable)

i_understand_unencrypted_risks: bool

Acknowledge unencrypted export risks (required in robot/JSON mode with –no-encryption)

preview: Option<PathBuf>

Preview an existing export locally (starts HTTP server)

port: u16

Port for preview server (default: 8080)

no_open: bool

Don’t auto-open browser when starting preview server

config: Option<String>

JSON config file for non-interactive export (use “-” for stdin)

validate_config: bool

Validate config file without running export

example_config: bool

Show example config file

Mirror(MirrorCommand)

Inspect and prune raw-mirror evidence under explicit operator control

Sources(SourcesCommand)

Manage remote sources (P5.x)

Models(ModelsCommand)

Manage semantic search models

Swarm(SwarmCommand)

Read-only swarm operations status

Import(ImportCommand)

Import data from external sources

Analytics(AnalyticsCommand)

Token usage, tool, and model analytics

Subcommands: status, tokens, tools, models, rebuild, validate. All subcommands share time-range, dimensional, and output flags.

Examples

cass analytics status --json
cass analytics tokens --days 7 --group-by day --json
cass analytics rebuild --json

Daemon

Run the semantic model daemon (Unix only)

Fields

socket: Option<PathBuf>

Socket path to listen on (default comes from env or built-in config)

idle_timeout: Option<u64>

Idle timeout in seconds before auto-shutdown (0 = never; default comes from env or built-in config)

max_connections: Option<usize>

Maximum concurrent connections (default comes from env or built-in config)

data_dir: Option<PathBuf>

Override data dir for model storage

Trait Implementations

impl Clone for Commands

fn clone(&self) -> Commands

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Commands

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl FromArgMatches for Commands

fn from_arg_matches(__clap_arg_matches: &ArgMatches) -> Result<Self, Error>

Instantiate Self from ArgMatches, parsing the arguments as needed.

fn from_arg_matches_mut( __clap_arg_matches: &mut ArgMatches, ) -> Result<Self, Error>

Instantiate Self from ArgMatches, parsing the arguments as needed.

fn update_from_arg_matches( &mut self, __clap_arg_matches: &ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

fn update_from_arg_matches_mut<'b>( &mut self, __clap_arg_matches: &mut ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

impl Subcommand for Commands

fn augment_subcommands<'b>(__clap_app: Command) -> Command

Append to Command so it can instantiate Self via FromArgMatches::from_arg_matches_mut

fn augment_subcommands_for_update<'b>(__clap_app: Command) -> Command

Append to Command so it can instantiate self via FromArgMatches::update_from_arg_matches_mut

fn has_subcommand(__clap_name: &str) -> bool

Test whether Self can parse a specific subcommand