Agent Bounties

Prepare bounty for wallet review

prepare_bounty_post

DestructiveIdempotent

Use this when the person's AI has gathered complete bounty terms and received explicit approval. Reuse posting_operation_id when resuming. For creator review, set review_mode=creator and the exact agreed delivery_deadline with timezone offset, omit automated benchmark/evidence fields, and explain that the creator confirms the verdict. Preserve parent bindings, the approved reference_attachment, and any approved bounty_image with its exact prompt and alt text; omit all three image fields together to use the deterministic review-page visual. This prepares a reviewable handoff; it does not publish, move funds, request a secret, or prove that a bounty exists.

Input Schema

TableJSON Schema

Name	Required	Description
`goal`	Yes	Public digital outcome the solver must deliver.
`title`	Yes	Concise public bounty title.
`benchmark`	No	Optional exact public sandboxed_regression_v1 benchmark. Supply this and evidence_schema together. The source commit, benchmark digest, OCI image, command, and resource limits are immutable bounty terms.
`crowdfund`	No	Keep false to fund on creation. Set true only to deposit 0 USDC now.
`meta_child`	No	Preserve the routed parent binding. The review page verifies the canonical parent before applying any child economics.
`source_url`	No	Optional public HTTPS source issue or task URL.
`review_mode`	No	Creator review commits the creator as the sole human reviewer. The creator confirms the verdict; AI approval has no settlement authority.
`bounty_image`	No	Optional approved image generated in a compatible AI conversation. Agent Bounties stores this exact file; omit it and both image text fields to use the deterministic fallback visual.
`image_prompt`	No	Optional exact prompt used to generate a user-approved bounty image. Supply this, image_alt_text, and bounty_image together, or omit all three to use the deterministic fallback visual.
`image_alt_text`	No	Optional accessible description of the approved image. Supply this, image_prompt, and bounty_image together.
`evidence_schema`	No	Optional public submission-evidence schema paired with benchmark. Supply both verifier fields or neither.
`discovery_source`	No	Optional public attribution for how the poster found Agent Bounties.
`task_window_days`	No	Optional bounded work window in days; defaults to 30.
`delivery_deadline`	No	For creator review, the exact agreed calendar deadline as an ISO timestamp with timezone offset. Never replace it with days after claim. Omit automated benchmark and evidence_schema fields in creator mode.
`solver_reward_usdc`	Yes	Solver reward in display USDC, for example 2.00.
`acceptance_criteria`	Yes	Binary or measurable public acceptance criteria.
`posting_operation_id`	No	Stable posting journey identifier. Reuse it when resuming or retrying the same operation. If omitted, a deterministic identifier is derived from the unchanged preparation arguments.
`reference_attachment`	No	Preserve the approved compact homepage-background snapshot descriptor, including its commit-pinned asset_url, sha256 and captured_at. The first-party review validates the pinned bytes; never replace the user's reference.
`verifier_reward_usdc`	Yes	Verifier reward and refundable claim bond in display USDC, for example 0.10. In creator mode, label this the creator-review reserve; it is paid to the creator on either verdict.

Output Schema

TableJSON Schema

Name	Required	Description	Default
`goal`	Yes
`image`	Yes
`state`	Yes
`title`	Yes
`schema`	Yes
`post_url`	Yes
`benchmark`	No
`crowdfund`	Yes
`meta_child`	No
`source_url`	No
`next_action`	Yes
`review_mode`	No
`target_usdc`	Yes
`bounty_created`	Yes
`evidence_schema`	No
`task_window_days`	Yes
`delivery_deadline`	No
`evidence_boundary`	Yes
`review_disclosure`	No
`solver_reward_usdc`	Yes
`acceptance_criteria`	Yes
`initial_funding_usdc`	Yes
`posting_operation_id`	No
`reference_attachment`	No
`verifier_reward_usdc`	Yes
`verification_prepared`	No
`wallet_signature_requested`	Yes

Schema Changelog

Changes observed during successful MCP inspections.

Changed15 schema fields changedSep 10, 2026

addedInput schema / properties / delivery_deadline

Added value: +{
+  "description": "For creator review, the exact agreed calendar deadline as an ISO timestamp with timezone offset. Never replace it with days after claim. Omit automated benchmark and evidence_schema fields in creator mode.",
+  "format": "date-time",
+  "type": [
+    "string",
+    "null"
+  ]
+}

addedInput schema / properties / meta_child

Added value: +{
+  "additionalProperties": false,
+  "description": "Preserve the routed parent binding. The review page verifies the canonical parent before applying any child economics.",
+  "properties": {
+    "intended_child_solver": {
+      "pattern": "^0x[0-9a-fA-F]{40}$",
+      "type": [
+        "string",
+        "null"
+      ]
+    },
+    "parent_bounty_contract": {
+      "pattern": "^0x[0-9a-fA-F]{40}$",
+      "type": "string"
+    }
+  },
+  "required": [
+    "parent_bounty_contract"
+  ],
+  "type": [
+    "object",
+    "null"
+  ]
+}

addedInput schema / properties / posting_operation_id

Added value: +{
+  "description": "Stable posting journey identifier. Reuse it when resuming or retrying the same operation. If omitted, a deterministic identifier is derived from the unchanged preparation arguments.",
+  "format": "uuid",
+  "type": [
+    "string",
+    "null"
+  ]
+}

addedInput schema / properties / reference_attachment

Added value: +{
+  "description": "Preserve the approved compact homepage-background snapshot descriptor, including its commit-pinned asset_url, sha256 and captured_at. The first-party review validates the pinned bytes; never replace the user's reference.",
+  "type": [
+    "object",
+    "null"
+  ]
+}

addedInput schema / properties / review_mode

Added value: +{
+  "description": "Creator review commits the creator as the sole human reviewer. The creator confirms the verdict; AI approval has no settlement authority.",
+  "enum": [
+    "creator",
+    "automated",
+    null
+  ],
+  "type": [
+    "string",
+    "null"
+  ]
+}

changedInput schema / properties / verifier_reward_usdc / description

Previous value: -"Verifier reward and refundable claim bond in display USDC, for example 0.10."New value: +"Verifier reward and refundable claim bond in display USDC, for example 0.10. In creator mode, label this the creator-review reserve; it is paid to the creator on either verdict."

addedOutput schema / properties / benchmark

Added value: +{
+  "type": [
+    "object",
+    "null"
+  ]
+}

addedOutput schema / properties / delivery_deadline

Added value: +{
+  "format": "date-time",
+  "type": [
+    "string",
+    "null"
+  ]
+}

addedOutput schema / properties / evidence_schema

Added value: +{
+  "type": [
+    "object",
+    "null"
+  ]
+}

addedOutput schema / properties / meta_child

Added value: +{
+  "type": [
+    "object",
+    "null"
+  ]
+}

addedOutput schema / properties / posting_operation_id

Added value: +{
+  "format": "uuid",
+  "type": "string"
+}

addedOutput schema / properties / reference_attachment

Added value: +{
+  "type": [
+    "object",
+    "null"
+  ]
+}

addedOutput schema / properties / review_disclosure

Added value: +{
+  "type": [
+    "string",
+    "null"
+  ]
+}

addedOutput schema / properties / review_mode

Added value: +{
+  "enum": [
+    "creator",
+    "automated"
+  ],
+  "type": "string"
+}

addedOutput schema / properties / verification_prepared
```
Added value: +{
+  "type": "boolean"
+}
```

Changed2 schema fields changedSep 5, 2026

addedInput schema / properties / benchmark

Added value: +{
+  "description": "Optional exact public sandboxed_regression_v1 benchmark. Supply this and evidence_schema together. The source commit, benchmark digest, OCI image, command, and resource limits are immutable bounty terms.",
+  "type": [
+    "object",
+    "null"
+  ]
+}

addedInput schema / properties / evidence_schema

Added value: +{
+  "description": "Optional public submission-evidence schema paired with benchmark. Supply both verifier fields or neither.",
+  "type": [
+    "object",
+    "null"
+  ]
+}

First observedSep 5, 2026

TDQS

A4.5/5.0

Behavior5/5

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

Even though annotations already mark the tool as non-read-only, idempotent, and destructive, the description adds valuable behavioral context: it is a preparation step, not a publishing/funding action; posting_operation_id should be reused when resuming; and creator review gives the creator settlement authority. It also explains the deterministic fallback visual and image-field coupling beyond what the annotations provide.

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?

The description is long but front-loaded with the trigger condition and packed with mode-specific rules. Nearly every sentence earns its place, though a few constraints, such as creator-mode omissions and image fallback behavior, are partially repeated from the schema.

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

Completeness5/5

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

For a 19-parameter tool with nested objects, an output schema, and two review modes, the description covers the key scenarios and invariants: resuming, creator review, parent preservation, image fallback, and the non-publishing boundary. An agent has enough context to invoke it correctly without additional discovery.

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

Parameters5/5

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

Schema coverage is 100%, but the description adds essential cross-parameter semantics: supply benchmark with evidence_schema, supply image fields together or omit all three, preserve parent bindings, and in creator mode omit automated fields and set a timezone-offset delivery_deadline. These grouping and mode-dependent rules are critical for constructing a valid call.

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

Purpose4/5

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

The description clearly identifies the verb and resource: prepare a bounty for wallet review after complete terms and explicit approval, and it closes with 'This prepares a reviewable handoff.' It does not explicitly name or contrast sibling tools such as prepare_bounty_action, so it stops just short of full sibling differentiation.

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

Usage Guidelines4/5

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

It gives an explicit precondition: use when the person's AI has gathered complete bounty terms and received explicit approval. It also clarifies what the tool does not do—publish, move funds, request a secret, or prove a bounty exists—but it does not explicitly route to an alternative sibling tool by name.

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

All Agent Bounties tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?