CtrlK
BlogDocsLog inGet started
Tessl Logo

cursor-agent-supervisor

Offloading tasks with a well-defined scope to sub-agents, for instance to use a sub-agent to implement a set of specs. Use this skill whenever a task should not need a broad knowledge of the whole project

76

1.59x
Quality

66%

Does it follow best practices?

Impact

91%

1.59x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./cursor-agent-supervisor/SKILL.md
SKILL.md
Quality
Evals
Security

Evaluation results

87%

72%

Automate Feature Branch Delegation

Sub-agent invocation and verification

Criteria
Without context
With context

Model discovery step

25%

100%

create-chat command

0%

100%

--resume command

0%

100%

--print flag present

0%

100%

Default model sonnet-4.5

0%

100%

Timeout 600000

10%

70%

Sequential two-step flow

0%

100%

Verification commands run

50%

100%

Diff or code review step

0%

0%

Failure handling

87%

100%

88%

6%

Draft Sub-Agent Task Brief for Payment Service Feature

Sub-agent task description quality

Criteria
Without context
With context

Setup section

37%

37%

Goal section

80%

90%

Key files section

50%

100%

Requirements section

87%

100%

Done when section

83%

91%

Specific goal, not vague

100%

100%

Concrete done criteria

90%

100%

No irrelevant context

87%

100%

No entire skill dumps

100%

100%

Skill or pattern guidance

87%

37%

Sub-agent scope is focused

100%

100%

100%

25%

Document Agent Supervision Protocol

Sub-agent recovery and supervision plan

Criteria
Without context
With context

Pre-delegation: environment setup

80%

100%

Sub-agent scope limited to coding

12%

100%

Post-completion: diff check

100%

100%

Post-completion: code review

90%

100%

Post-completion: verification commands

100%

100%

Recovery: agent failure

50%

100%

Recovery: incomplete work

100%

100%

Recovery: off-track implementation

90%

100%

Recovery: tooling issues

20%

100%

Actionable and ordered steps

91%

100%

Repository
YPares/agent-skills
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.