CtrlK
BlogDocsLog inGet started
Tessl Logo

coding-agent-helpers/compact-handoff

Use when work needs to be handed off to another agent or another human. Produce a continuation-ready brief with the objective, completed work, assumptions, unresolved issues, and next action instead of a generic summary. Good triggers include "prepare a handoff", "make this resumable", and "summarize this for another agent".

92

1.41x
Quality

100%

Does it follow best practices?

Impact

89%

1.41x

Average score across 8 eval scenarios

SecuritybySnyk

Passed

No known issues

Overview
Quality
Evals
Security
Files

criteria.jsonevals/scenario-7/

{
  "context": "The agent must produce a handoff for a cloud infrastructure setup with multiple open questions and a critical blocker. This scenario tests whether the agent surfaces unresolved issues and dependencies clearly, and whether the output is continuation-ready rather than a generic project status report.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Unresolved section present",
      "description": "The document has a section explicitly labelled 'Unresolved' (or equivalent) containing open questions or blockers.",
      "max_score": 8
    },
    {
      "name": "Terraform state blocker in Unresolved",
      "description": "The Unresolved section explicitly flags the Terraform state migration as a blocker that must be resolved before running terraform apply.",
      "max_score": 12
    },
    {
      "name": "Open decisions in Unresolved",
      "description": "The Unresolved section contains at least one genuine decision question (e.g. Multi-AZ choice or SQS queue dependency), not just a list of remaining tasks.",
      "max_score": 10
    },
    {
      "name": "Next Action addresses the blocker",
      "description": "The Next Action section names the Terraform state migration (or another highest-priority unblocking step) as the first action.",
      "max_score": 10
    },
    {
      "name": "Next Action is specific",
      "description": "The Next Action contains a specific command, file path, or resource name — not a vague instruction like 'continue setup'.",
      "max_score": 8
    },
    {
      "name": "Critical References section present",
      "description": "The document has a Critical References section listing at least one concrete identifier (account ID, cluster name, path, or ARN).",
      "max_score": 8
    },
    {
      "name": "Critical References are minimal",
      "description": "Critical References does NOT list every single resource mentioned in the context; it surfaces only those needed for immediate continuity.",
      "max_score": 8
    },
    {
      "name": "Continuation-ready tone",
      "description": "The document is written for action, not record-keeping: it does NOT read as a project status report or completion certificate.",
      "max_score": 8
    },
    {
      "name": "All six sections present",
      "description": "The document includes all six prescribed sections: Objective, Completed, Assumptions, Unresolved, Next Action, Critical References.",
      "max_score": 8
    },
    {
      "name": "Assumptions section non-empty",
      "description": "The Assumptions section contains at least one stated assumption (e.g. about environment, access, or decisions treated as settled).",
      "max_score": 10
    },
    {
      "name": "No narrative recap",
      "description": "The document does not contain a prose narrative of what the contractor did; completed work is listed, not explained.",
      "max_score": 10
    }
  ]
}

evals

tile.json