Skip to main content
Glama

submit_and_pay_job_from_file

Same as submit_and_pay_job, but the job input is read from a file on disk by the MCP server instead of being passed inline by the LLM.

Instructions

Same as submit_and_pay_job, but the job input is read from a file on disk by the MCP server instead of being passed inline by the LLM. Use this when the input is large or binary (images, logs, captured output) and the LLM only needs to forward it - the file content never enters the model's output tokens. input_path may be absolute or relative to the MCP server's working directory. The file is ALWAYS transferred peer-to-peer via iroh, so this needs: a persistent agent, a PAID provider skill (free skills reject file inputs), and the iroh addon. Text files reach the skill on stdin; binary files via ELISYM_INPUT_FILE. Pass an optional prompt to send a text instruction alongside the file (e.g. how to edit an image); it rides inline (encrypted) while the file rides P2P. This pays the listed price up front; when the capability advertises delegation, use submit_delegated_job_from_file so a metered card bills only actual usage.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "input_path": "<input_path>",
      "provider_npub": "<provider_npub>"
    },
    "name": "submit_and_pay_job_from_file"
  }
}

Input Schema

NameRequiredDescriptionDefault
promptNoOptional text instruction sent alongside the file (e.g. how to edit an image: "make it night", "add a hat"). It rides inline (NIP-44 encrypted) in the job event while the file travels peer-to-peer via iroh. The single attachment slot holds the file, so the prompt cannot spill to a second transfer - keep it short.
capabilityNogeneral
input_pathYesPath to a regular file whose contents become the job input. Absolute or relative to the MCP server's working directory.
session_idNoConversation control. Omit for automatic session management (providers advertising context support get a conversation auto-started on first contact; an ongoing conversation triggers a continue/new/one-off question before anything is published). Pass "new" to force a fresh conversation, "none" to force a stateless one-off, or a session_id from a previous result to continue that conversation. The provider answers with the conversation context of prior exchanges under the same id.
kind_offsetNo
timeout_secsNo
provider_npubYes
allow_outside_cwdNoAllow reading a file outside the MCP server working directory. Off by default - the file content is forwarded to the provider before payment and is invisible in the transcript, so reads are confined to the working dir unless this is set. Sensitive files (secret keys, .env, SSH/keypair, ~/.elisym, /proc) are always refused.
max_price_lamportsNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed2 schema fields changedv0.1.60
    • addedInput schema / properties / max_price_lamports / maximum
      Added value: +9007199254740991
    • addedInput schema / properties / max_price_lamports / minimum
      Added value: +-9007199254740991
  2. Changed2 schema fields changedv0.1.42
    • addedInput schema / properties / prompt
      Added value: +{
      +  "default": "",
      +  "description": "Optional text instruction sent alongside the file (e.g. how to edit an image: \"make it night\", \"add a hat\"). It rides inline (NIP-44 encrypted) in the job event while the file travels peer-to-peer via iroh. The single attachment slot holds the file, so the prompt cannot spill to a second transfer - keep it short.",
      +  "maxLength": 100000,
      +  "type": "string"
      +}
    • addedInput schema / properties / session_id
      Added value: +{
      +  "anyOf": [
      +    {
      +      "const": "new",
      +      "type": "string"
      +    },
      +    {
      +      "const": "none",
      +      "type": "string"
      +    },
      +    {
      +      "pattern": "^[0-9a-f]{8}-[0-9a-f]{4}-4[0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$",
      +      "type": "string"
      +    }
      +  ],
      +  "description": "Conversation control. Omit for automatic session management (providers advertising context support get a conversation auto-started on first contact; an ongoing conversation triggers a continue/new/one-off question before anything is published). Pass \"new\" to force a fresh conversation, \"none\" to force a stateless one-off, or a session_id from a previous result to continue that conversation. The provider answers with the conversation context of prior exchanges under the same id."
      +}
  3. Addedv0.1.37
  4. Removedv0.1.36
  5. Addedv0.1.2

TDQS

A4.5/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are supplied, so the description carries the full disclosure burden and does so unusually well: it explains the P2P iroh transfer path, that the file content never enters the model's output tokens, how text goes to stdin while binary goes via ELISYM_INPUT_FILE, that payment (the listed price) happens up front before the job, and that out-of-working-dir reads are off by default with sensitive paths always refused. These are non-obvious risks and preconditions an agent cannot infer from the schema.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the differentiator and the use-when rule in the first two sentences, then flowing into prerequisites, transfer mechanics, and the delegation alternative. It is dense but nearly every clause carries decision-relevant information; the prompt example slightly overlaps the schema description and the sentences run long, which keeps it from a 5.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 9-parameter, no-annotation, no-output-schema mutation-plus-payment tool, the definition covers safety, transfer path, cost model, and the sibling alternative, which is close to complete. The remaining gap is the undocumented knobs (capability, timeout_secs, kind_offset, max_price_lamports) that an agent must call blind.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is only 44%, so the description must compensate, and it does for the highest-risk parameters: input_path (absolute or relative to the server cwd), prompt (inline encrypted, single attachment slot, keep it short), and allow_outside_cwd (confined by default, secrets refused). However capability, kind_offset, timeout_secs, provider_npub, and max_price_lamports get no meaning from either the description or the schema, leaving nearly half the surface undefined.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource (submit and pay a job) and immediately pins its distinctive scope: the input is read from a file on disk by the MCP server rather than passed inline. It also names the sibling it is a variant of (submit_and_pay_job), so an agent can pick between the two without opening either schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives an explicit when-to-use rule ('input is large or binary... the LLM only needs to forward it') and, crucially, a when-not/alternative rule routed to a named sibling: use submit_delegated_job_from_file when the capability advertises delegation so a metered card bills only actual usage. It also lists selection-relevant prerequisites (persistent agent, PAID provider skill, iroh addon) that rule the tool in or out for a given setup.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.