CtrlK
BlogDocsLog inGet started
Tessl Logo

my-custom-skill

Example custom agent skill — a template to get started.

61

2.41x
Quality

46%

Does it follow best practices?

Impact

82%

2.41x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./packages/agentsociety2/agentsociety2/custom/skills/examples/my-custom-skill/SKILL.md
SKILL.md
Quality
Evals
Security

Evaluation results

100%

60%

Build a Habit Tracking Skill

Pattern A skill authoring workflow

Criteria
Without this skill
With this skill

No script field

100%

100%

ask_env referenced

0%

100%

ask_env is first

0%

100%

read referenced

50%

100%

write or append referenced

50%

100%

finish referenced

0%

100%

finish is last

0%

100%

finish with summary

0%

100%

No execute_skill_script

100%

100%

No non-atomic tools

100%

100%

Correct step order

50%

100%

100%

67%

Build Two Skills for the Engineering Team's Workflow Automation Hub

Pattern A vs Pattern B skill discrimination

Criteria
Without this skill
With this skill

No script in meeting-summarizer

100%

100%

Script field in checksum-verifier

0%

100%

meeting-summarizer uses ask_env

0%

100%

meeting-summarizer uses read/write/append

30%

100%

meeting-summarizer calls finish

0%

100%

checksum script defines entrypoint

0%

100%

checksum script outputs JSON to stdout

100%

100%

checksum script parses --args-json

0%

100%

checksum-verifier uses execute_skill_script

0%

100%

No LLM logic in checksum script

100%

100%

47%

17%

Workspace Text Analysis Skill

Pattern B script-based skill creation

Criteria
Without this skill
With this skill

Script field in frontmatter

0%

0%

entrypoint function defined

0%

0%

args-json argument parsing

0%

0%

JSON output to stdout

0%

0%

Path.cwd() for workspace I/O

100%

100%

No LLM access

100%

100%

execute_skill_script in SKILL.md

0%

70%

Environment variable awareness

0%

100%

Deterministic txt file computation

100%

100%

Repository
tsinghua-fib-lab/AgentSociety
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.