Skip to main content
Glama

update_dependency

Destructive

Sets the condition, criterion, or verifiable evidence for an existing dependency edge so the dependent task is released only when Mesh can confirm the blocker meets it.

Instructions

Set the condition, criterion and/or evidence of an existing dependency edge (task_id depends on depends_on_task_id). Evidence releases the dependent only when Mesh can verify it: send evidence_criterion equal to the edge criterion (default: the condition name). shipped/deployed/accepted are checked against the blocker's own pr link: head_sha must be the link's current HEAD (and environment, e.g. prod, for deployed); evidence_url is only stored, never fetched. access_ready/contract_ready/artifact_published need attest=true from someone with access to the blocker. A bare evidence_url is kept as an unverified link and releases nothing. Changing condition or criterion needs expected_generation: the generation you read from the edge (get_task dependencies) — missing is refused, stale is a 409. An empty evidence_url re-arms the edge.

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": {
      "task_id": "<task_id>",
      "depends_on_task_id": "<depends_on_task_id>"
    },
    "name": "update_dependency"
  }
}

Input Schema

NameRequiredDescriptionDefault
attestNoaccess_ready/contract_ready/artifact_published: explicit attestation by you that the criterion holds.
task_idYesThe blocked task (the one that has the dependency).
head_shaNoshipped/deployed/accepted: the blocker pr link's current HEAD the receipt covers.
conditionNoWhat a blocks edge waits for. task_done (default): the blocker reaches done or cancelled. access_ready / contract_ready / artifact_published: released by evidence_url, or when the blocker reaches done or cancelled. shipped: the blocker is marked shipped or reaches done (cancelled does not release it). deployed: the blocker is marked shipped or a VCS link of it has a recorded deploy (see mark_vcs_deployed); merged/done alone does not release it. accepted: the blocker reaches done (cancelled does not release it). Evidence (evidence_url) releases an edge of any condition.
criterionNoWhat this edge waits for, in words (default: the condition name). Evidence must name it. Changing it needs expected_generation.
environmentNodeployed: the environment of the deploy receipt, e.g. prod.
evidence_urlNoReference to the proof; stored, never fetched. Empty string clears the evidence.
depends_on_task_idYesThe blocker the edge points at.
evidence_criterionNoThe criterion the evidence is for; must equal the edge criterion or the evidence stays unverified/invalid.
expected_generationNoThe edge generation you saw. Required to change condition or criterion; if sent with evidence a stale value is rejected (409).

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed7 schema fields changedv0.1.7
    • addedInput schema / properties / attest
      Added value: +{
      +  "description": "access_ready/contract_ready/artifact_published: explicit attestation by you that the criterion holds.",
      +  "type": "boolean"
      +}
    • addedInput schema / properties / criterion
      Added value: +{
      +  "description": "What this edge waits for, in words (default: the condition name). Evidence must name it. Changing it needs expected_generation.",
      +  "type": "string"
      +}
    • addedInput schema / properties / environment
      Added value: +{
      +  "description": "deployed: the environment of the deploy receipt, e.g. prod.",
      +  "type": "string"
      +}
    • addedInput schema / properties / evidence_criterion
      Added value: +{
      +  "description": "The criterion the evidence is for; must equal the edge criterion or the evidence stays unverified/invalid.",
      +  "type": "string"
      +}
    • changedInput schema / properties / evidence_url / description
      Previous value: -"Proof the condition holds; empty string clears it."New value: +"Reference to the proof; stored, never fetched. Empty string clears the evidence."
    • addedInput schema / properties / expected_generation
      Added value: +{
      +  "description": "The edge generation you saw. Required to change condition or criterion; if sent with evidence a stale value is rejected (409).",
      +  "type": "number"
      +}
    • addedInput schema / properties / head_sha
      Added value: +{
      +  "description": "shipped/deployed/accepted: the blocker pr link's current HEAD the receipt covers.",
      +  "type": "string"
      +}
  2. Addedv0.1.6

TDQS

A4.3/5.0
Behavior5/5

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

Annotations only flag a destructive, non-idempotent write; the description adds the operationally important details: evidence_url is stored but never fetched, a bare URL releases nothing, a missing expected_generation is refused while a stale one returns 409, and an empty evidence_url re-arms the edge. That is real error, auth and side-effect disclosure beyond the structured hints.

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?

Purpose and edge identity are front-loaded, then verification rules follow in a logical order. It is dense and the 'evidence_url is only stored' and 'bare evidence_url releases nothing' sentences partially overlap, costing some tightness.

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?

Ten parameters, no output schema, and non-obvious coupling between condition, criterion and evidence are all addressed, including the generation-based concurrency rule. The remaining unstated item (the exact response shape) is not required because no output schema exists, and callers can proceed correctly from this text.

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?

Schema coverage is 100%, so baseline is 3, but the description adds relationships the schema does not spell out in one place: evidence_criterion must match the edge criterion (default = condition name), head_sha must be the link's current HEAD, and environment applies only to deployed. A few parameters (attest, expected_generation) are only restated from the schema.

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?

Opens with a specific verb and resource: 'Set the condition, criterion and/or evidence of an existing dependency edge', and pins the edge identity to (task_id, depends_on_task_id). The word 'existing' implicitly separates it from add_dependency/remove_dependency, but no sibling is named, so the differentiation is inferential rather than explicit.

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?

Gives rich conditional guidance: evidence releases only when Mesh can verify it, shipped/deployed/accepted need head_sha from the blocker's pr link, and ready-type criteria need attest=true from someone with blocker access. It also points to get_task dependencies as the way to read the generation. It never states when to prefer add_dependency/remove_dependency instead, which keeps it below a 5.

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