dsh-relay

Overview Schema Source Related Servers Discussions Admin

Dispatch a read-only Harness review

start_review

Launch a read-only Harness code review session from workspace paths, returning a stable link your agent can monitor until the run finishes.

Instructions

Create or reuse a native Harness session with the permission preset fixed to read-only, then return a stable session link. Task parameters are path-reference-only: provide the authorized workspace, reviewTargets, contextReadScope, excludedPaths, writeScope, and acceptance criteria. reviewTargets identify what to assess; they are not a read whitelist. contextReadScope declares where Harness may search and read supporting implementation, tests, configuration, and architecture material. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. These fields are task instructions, not enforced per-path filesystem isolation. The selected Harness model may send content it reads to its configured model provider even though Relay itself uses a loopback address. After start succeeds, share webUrl and keep wait_run until succeeded/failed/cancelled/needs_attention. The calling agent MUST read assistantText before claiming the review is done. If the parent user asked to review then fix, apply accepted findings only after the run is terminal; do not treat a still-running review as finished.

Input Schema

TableJSON Schema

Name	Required	Description
`task`	No	Task parameters are path-reference-only: provide the authorized workspace, reviewTargets, contextReadScope, excludedPaths, writeScope, and acceptance criteria. reviewTargets identify what to assess; they are not a read whitelist. contextReadScope declares where Harness may search and read supporting implementation, tests, configuration, and architecture material. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. These fields are task instructions, not enforced per-path filesystem isolation. The selected Harness model may send content it reads to its configured model provider even though Relay itself uses a loopback address.
`model`	No
`content`	No	Task parameters are path-reference-only: provide the authorized workspace, reviewTargets, contextReadScope, excludedPaths, writeScope, and acceptance criteria. reviewTargets identify what to assess; they are not a read whitelist. contextReadScope declares where Harness may search and read supporting implementation, tests, configuration, and architecture material. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. These fields are task instructions, not enforced per-path filesystem isolation. The selected Harness model may send content it reads to its configured model provider even though Relay itself uses a loopback address.
`provider`	No
`sessionId`	No
`workspace`	Yes	Authorized absolute Harness workspace root; Harness reads named in-scope files from here.
`writeScope`	No	Paths Harness may modify. Must be empty for read-only runs; this is an instruction declaration in addition to native permissions.
`agentPreset`	No
`openBrowser`	No	Keep false unless the user explicitly asks to open the Harness page in the OS browser.
`sessionMode`	No
`excludedPaths`	No	Paths excluded from contextual reading. This is an instruction declaration, not filesystem enforcement.
`reviewTargets`	No	Subjects to assess. This is not a read whitelist; supply together with contextReadScope.
`idempotencyKey`	No
`reasoningEffort`	No
`contextReadScope`	No	Workspace paths Harness may search and read for evidence. Use only the targets when the user explicitly requests a target-only review.
`authorizationBasis`	No	Truthful evidence that the user explicitly selected Harness to inspect this workspace. Set only when the current conversation contains that instruction. This records evidence and does not expand permissions or guarantee approval.

Schema Changelog

Changes observed during successful MCP inspections.

Changed9 schema fields changedv0.2.13Sep 10, 2026

addedInput schema / properties / authorizationBasis

Added value: +{
+  "const": "explicit-user-request",
+  "description": "Truthful evidence that the user explicitly selected Harness to inspect this workspace. Set only when the current conversation contains that instruction. This records evidence and does not expand permissions or guarantee approval.",
+  "type": "string"
+}

changedInput schema / properties / content / description

Previous value: -"Task parameters are path-reference-only: provide the authorized workspace, file or directory locations, scope, and acceptance criteria. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. This rule is identical for read-only and write-capable permissions."New value: +"Task parameters are path-reference-only: provide the authorized workspace, reviewTargets, contextReadScope, excludedPaths, writeScope, and acceptance criteria. reviewTargets identify what to assess; they are not a read whitelist. contextReadScope declares where Harness may search and read supporting implementation, tests, configuration, and architecture material. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. These fields are task instructions, not enforced per-path filesystem isolation. The selected Harness model may send content it reads to its configured model provider even though Relay itself uses a loopback address."

changedInput schema / properties / content / items / oneOf

Previous value: -[
-  {
-    "properties": {
-      "text": {
-        "description": "Task parameters are path-reference-only: provide the authorized workspace, file or directory locations, scope, and acceptance criteria. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. This rule is identical for read-only and write-capable permissions.",
-        "maxLength": 100000,
-        "type": "string"
-      },
-      "type": {
-        "const": "text",
-        "type": "string"
-      }
-    },
-    "required": [
-      "type",
-      "text"
-    ],
-    "type": "object"
-  },
-  {
-    "properties": {
-      "data": {
-        "minLength": 1,
-        "type": "string"
-      },
-      "mediaType": {
-        "enum": [
-          "image/png",
-          "image/jpeg",
-          "image/webp",
-          "image/gif"
-        ],
-        "type": "string"
-      },
-      "name": {
-        "type": "string"
-      },
-      "type": {
-        "const": "image",
-        "type": "string"
-      }
-    },
-    "required": [
-      "type",
-      "mediaType",
-      "data"
-    ],
-    "type": "object"
-  }
-]New value: +[
+  {
+    "properties": {
+      "text": {
+        "description": "Task parameters are path-reference-only: provide the authorized workspace, reviewTargets, contextReadScope, excludedPaths, writeScope, and acceptance criteria. reviewTargets identify what to assess; they are not a read whitelist. contextReadScope declares where Harness may search and read supporting implementation, tests, configuration, and architecture material. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. These fields are task instructions, not enforced per-path filesystem isolation. The selected Harness model may send content it reads to its configured model provider even though Relay itself uses a loopback address.",
+        "maxLength": 100000,
+        "type": "string"
+      },
+      "type": {
+        "const": "text",
+        "type": "string"
+      }
+    },
+    "required": [
+      "type",
+      "text"
+    ],
+    "type": "object"
+  },
+  {
+    "properties": {
+      "data": {
+        "minLength": 1,
+        "type": "string"
+      },
+      "mediaType": {
+        "enum": [
+          "image/png",
+          "image/jpeg",
+          "image/webp",
+          "image/gif"
+        ],
+        "type": "string"
+      },
+      "name": {
+        "type": "string"
+      },
+      "type": {
+        "const": "image",
+        "type": "string"
+      }
+    },
+    "required": [
+      "type",
+      "mediaType",
+      "data"
+    ],
+    "type": "object"
+  }
+]

addedInput schema / properties / contextReadScope

Added value: +{
+  "description": "Workspace paths Harness may search and read for evidence. Use only the targets when the user explicitly requests a target-only review.",
+  "items": {
+    "description": "Workspace-relative path, or an absolute path contained by the authorized workspace.",
+    "minLength": 1,
+    "type": "string"
+  },
+  "minItems": 1,
+  "type": "array"
+}

addedInput schema / properties / excludedPaths

Added value: +{
+  "description": "Paths excluded from contextual reading. This is an instruction declaration, not filesystem enforcement.",
+  "items": {
+    "description": "Workspace-relative path, or an absolute path contained by the authorized workspace.",
+    "minLength": 1,
+    "type": "string"
+  },
+  "type": "array"
+}

addedInput schema / properties / openBrowser / description

Added value: +"Keep false unless the user explicitly asks to open the Harness page in the OS browser."

addedInput schema / properties / reviewTargets

Added value: +{
+  "description": "Subjects to assess. This is not a read whitelist; supply together with contextReadScope.",
+  "items": {
+    "description": "Workspace-relative path, or an absolute path contained by the authorized workspace.",
+    "minLength": 1,
+    "type": "string"
+  },
+  "minItems": 1,
+  "type": "array"
+}

changedInput schema / properties / task / description

Previous value: -"Task parameters are path-reference-only: provide the authorized workspace, file or directory locations, scope, and acceptance criteria. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. This rule is identical for read-only and write-capable permissions."New value: +"Task parameters are path-reference-only: provide the authorized workspace, reviewTargets, contextReadScope, excludedPaths, writeScope, and acceptance criteria. reviewTargets identify what to assess; they are not a read whitelist. contextReadScope declares where Harness may search and read supporting implementation, tests, configuration, and architecture material. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. These fields are task instructions, not enforced per-path filesystem isolation. The selected Harness model may send content it reads to its configured model provider even though Relay itself uses a loopback address."

addedInput schema / properties / writeScope

Added value: +{
+  "description": "Paths Harness may modify. Must be empty for read-only runs; this is an instruction declaration in addition to native permissions.",
+  "items": {
+    "description": "Workspace-relative path, or an absolute path contained by the authorized workspace.",
+    "minLength": 1,
+    "type": "string"
+  },
+  "type": "array"
+}

Changed4 schema fields changedv0.2.9Sep 1, 2026

addedInput schema / properties / content / description

Added value: +"Task parameters are path-reference-only: provide the authorized workspace, file or directory locations, scope, and acceptance criteria. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. This rule is identical for read-only and write-capable permissions."

changedInput schema / properties / content / items / oneOf

Previous value: -[
-  {
-    "properties": {
-      "text": {
-        "maxLength": 100000,
-        "type": "string"
-      },
-      "type": {
-        "const": "text",
-        "type": "string"
-      }
-    },
-    "required": [
-      "type",
-      "text"
-    ],
-    "type": "object"
-  },
-  {
-    "properties": {
-      "data": {
-        "minLength": 1,
-        "type": "string"
-      },
-      "mediaType": {
-        "enum": [
-          "image/png",
-          "image/jpeg",
-          "image/webp",
-          "image/gif"
-        ],
-        "type": "string"
-      },
-      "name": {
-        "type": "string"
-      },
-      "type": {
-        "const": "image",
-        "type": "string"
-      }
-    },
-    "required": [
-      "type",
-      "mediaType",
-      "data"
-    ],
-    "type": "object"
-  }
-]New value: +[
+  {
+    "properties": {
+      "text": {
+        "description": "Task parameters are path-reference-only: provide the authorized workspace, file or directory locations, scope, and acceptance criteria. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. This rule is identical for read-only and write-capable permissions.",
+        "maxLength": 100000,
+        "type": "string"
+      },
+      "type": {
+        "const": "text",
+        "type": "string"
+      }
+    },
+    "required": [
+      "type",
+      "text"
+    ],
+    "type": "object"
+  },
+  {
+    "properties": {
+      "data": {
+        "minLength": 1,
+        "type": "string"
+      },
+      "mediaType": {
+        "enum": [
+          "image/png",
+          "image/jpeg",
+          "image/webp",
+          "image/gif"
+        ],
+        "type": "string"
+      },
+      "name": {
+        "type": "string"
+      },
+      "type": {
+        "const": "image",
+        "type": "string"
+      }
+    },
+    "required": [
+      "type",
+      "mediaType",
+      "data"
+    ],
+    "type": "object"
+  }
+]

addedInput schema / properties / task / description

Added value: +"Task parameters are path-reference-only: provide the authorized workspace, file or directory locations, scope, and acceptance criteria. Never embed source text, diffs, file dumps, encoded source, or repository archives. Harness reads named files from the authorized workspace itself. This rule is identical for read-only and write-capable permissions."

addedInput schema / properties / workspace / description

Added value: +"Authorized absolute Harness workspace root; Harness reads named in-scope files from here."

First observedv0.2.6Aug 30, 2026

TDQS

A4.4/5.0

Behavior5/5

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

The description goes beyond annotations by disclosing that the Harness model may send content to its configured provider despite Relay using a loopback address. It also clarifies that the task fields are 'task instructions, not enforced per-path filesystem isolation,' which is a significant behavioral caveat. This is valuable context that annotations (openWorldHint, destructiveHint) do not cover. No contradiction with annotations.

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

Conciseness3/5

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

The description is comprehensive but quite long and repetitive. It repeats the same 'path-reference-only' guidance multiple times and includes schema-like details that are already in the input schema. While it's structured with clear points, it could be more concise by not duplicating parameter descriptions that are already present in 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?

Given the tool's complexity (16 parameters, high-stakes review workflow), the description covers all critical aspects: what to provide, how to handle results, and caveats about read-only and data transmission. It explains the full lifecycle from starting the review to waiting for completion and handling fixes. No output schema exists, so the description's emphasis on sharing webUrl and reading assistantText fills that gap.

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

Parameters4/5

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

The description and schema already provide extensive parameter documentation for reviewTargets, contextReadScope, excludedPaths, and writeScope. The description reinforces that reviewTargets is not a read whitelist and that contextReadScope declares where Harness can search. Many parameters (model, provider, sessionMode, reasoningEffort, agentPreset) have no description in the schema, but the description covers the critical ones for correct usage, so it compensates for the 56% coverage.

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 states the tool creates or reuses a native Harness session with read-only permissions and returns a session link. It distinguishes itself from siblings like start_run (generic runs) and start_service (start a service) by focusing on reviews. However, it could more explicitly contrast with start_run to clarify when a review session differs from a generic run.

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?

The description provides explicit guidance on when to use this tool (for read-only reviews), how to structure the task parameters (path-reference-only), and what not to do (never embed source text). It also states that after starting, the agent must share webUrl and keep wait_run until terminal, and must read assistantText before claiming completion. It even addresses the 'review then fix' scenario, instructing to apply findings only after the run is terminal.

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

Dispatch a read-only Harness review

Instructions

Input Schema

Schema Changelog

TDQS

Other Tools