> ## Documentation Index > Fetch the complete documentation index at: https://docs.crewship.ai/llms.txt > Use this file to discover all available pages before exploring further. # crewship context > Inspect authorized restricted context, immutable project files and private routine results. # crewship context These commands use the authenticated server API and your selected workspace. Use exact chat, agent, project, file, version and run IDs; commands do not read another agent's directory or a local database. Server audience and resource grants apply on every request. Lists and receipts print JSON; downloads print bytes and `ask` prints the server's SSE stream. ## chat inspect ```bash theme={null} crewship context chat inspect ``` Read your authorized context history, summaries and memory notes. ## chat memory ```bash theme={null} crewship context chat memory ``` Read only the chat's authorized memory notes. ## chat memory-add ```bash theme={null} crewship context chat memory-add "Remember this preference" ``` Add a private classified note of at most 8192 bytes. ## chat memory-delete ```bash theme={null} crewship context chat memory-delete ``` Delete an exact authorized note. Existing dependent attempts are revoked by the server. ## chat files ```bash theme={null} crewship context chat files ``` List immutable output file metadata for your authorized chat audience. ## chat download ```bash theme={null} crewship context chat download --out result.txt ``` | Flag | Default | Description | | - | - | - | | `--out` | `-` | Exact local destination, or `-` for stdout. File downloads replace the destination only after a complete successful transfer. | ## chat attempts ```bash theme={null} crewship context chat attempts ``` Inspect your visible attempt outcomes without exposing opaque execution handles. ## chat profile ```bash theme={null} crewship context chat profile ``` Inspect the server-selected chat execution profile. ## chat ask ```bash theme={null} crewship context chat ask "Explain the current result" ``` Send an ordinary restricted chat turn, requiring chat rights. The server chooses the installed profile, model, scoped prompt and provider; the command prints SSE frames. For an explicit host run requiring run rights, use `crewship run` and its declared project file version options. ## execution-profile get ```bash theme={null} crewship context execution-profile get ``` Read the agent's restricted profile with agent edit authorization. ## execution-profile set ```bash theme={null} crewship context execution-profile set native_api_key ``` Set `disabled`, `responses_text` or `native_api_key`. The server revokes existing bound attempts when the profile changes. Enabling a profile does not install its runtime or bypass admission, model, credential or budget checks. ## project-files list ```bash theme={null} crewship context project-files list ``` List current immutable version IDs and revision metadata with project read authorization. ## project-files upload ```bash theme={null} crewship context project-files upload ./input.txt --name docs/input.txt crewship context project-files upload ./new.txt --file-id --expected-revision 3 ``` Upload a bounded regular local file as base64 bytes. Replacements require the exact file ID and current revision; a concurrent revision change returns a conflict. | Flag | Default | Description | | - | - | - | | `--name` | Local basename | Project-relative file name. | | `--file-id` | Empty | Existing file ID to replace; empty creates a new file. | | `--expected-revision` | `0` | Exact positive current revision for replacement; zero only creates a file. | ## project-files retire ```bash theme={null} crewship context project-files retire --expected-revision 3 ``` Retire the exact current file with project write authorization. The server revokes dependent execution attempts. | Flag | Default | Description | | - | - | - | | `--expected-revision` | `0` | Required positive current file revision. | ## project-files download ```bash theme={null} crewship context project-files download --out input.txt ``` Download the exact immutable authorized version, without substituting a current version. | Flag | Default | Description | | - | - | - | | `--out` | `-` | Exact local destination, or `-` for stdout; local replacement occurs only after successful transfer. | ## issue-preflight ```bash theme={null} crewship context issue-preflight --claim --agent-id --routine-slug ``` Explicitly claim an assigned project-backed issue and enqueue its declared private preflight. The server selects the issue brief and actor, checks assignment, project rights and budget, and returns a private receipt. This is a mutation and never runs as an automatic list/read side effect. | Flag | Default | Description | | - | - | - | | `--claim` | `false` | Required explicit authorization to claim and enqueue. | | `--agent-id` | Empty | Required exact assigned agent ID. | | `--routine-slug` | Empty | Required declared private routine slug. | ## routine-runs list ```bash theme={null} crewship context routine-runs list ``` List your authorized private run receipts in the selected workspace. ## routine-runs result ```bash theme={null} crewship context routine-runs result ``` Read an exact authorized private receipt and current classified output proof. Revoked source or provider authority prevents reading stale output. ## Authorized routine catalog `crewship context routines` lists the server-authorized private routine catalog for the selected workspace, with declared scalar input names and frozen definition/execution hashes. It returns no prompts, provider credentials or denied routines. ## chat project-input-options ```bash theme={null} crewship context chat project-input-options --search invoice ``` List at most 100 currently authorized native project inputs, without source bytes. An empty search lists the available metadata; narrow it when `has_more` is true. | Flag | Default | Description | | - | - | - | | `--search` | empty | Filename or project-name substring, at most 128 bytes. | ## Authorized Pages `crewship context pages` lists only Pages with currently executable declared actions in the selected workspace. The authenticated projection returns current publication identifiers and scalar input descriptors; it excludes fixed parameters, input defaults and producer data. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.