CtrlK
BlogDocsLog inGet started
Tessl Logo

autonomous-agents

Autonomous agents are AI systems that can independently decompose goals, plan actions, execute tools, and self-correct without constant human guidance. The challenge isn't making them capable - it'...

49

1.18x

Quality

27%

Does it follow best practices?

Impact

82%

1.18x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./skills/antigravity-autonomous-agents/SKILL.md
SKILL.md
Quality
Evals
Security

Evaluation results

100%

23%

Personal Finance Monitoring Agent

Guardrails and cost controls

Criteria
Without context
With context

Constrained scope

100%

100%

Hard cost limit

25%

100%

Least privilege access

100%

100%

No general-purpose autonomy

100%

100%

Structured logging

100%

100%

Observability in design

100%

100%

Hard action limit

55%

100%

Guardrails described

88%

100%

Output validation

50%

100%

Incremental autonomy note

37%

100%

Without context: $0.6309 · 3m 4s · 17 turns · 16 in / 11,803 out tokens

With context: $0.6697 · 3m 10s · 23 turns · 23 in / 11,330 out tokens

68%

21%

Research Paper Metadata Extraction Agent

ReAct and Plan-Execute patterns

Criteria
Without context
With context

Separate planning phase

8%

0%

Execution separate from planning

10%

0%

ReAct-style loop

83%

83%

Step count bounded

0%

100%

Trace log generated

100%

100%

Ground truth validation

66%

100%

Corrupted paper handled

100%

100%

Robust error handling

20%

70%

Structured step logging

62%

75%

Compounding failure awareness

25%

62%

Without context: $0.6103 · 2m 48s · 24 turns · 71 in / 8,107 out tokens

With context: $1.1932 · 4m 44s · 37 turns · 83 in / 17,048 out tokens

79%

-6%

Blog Post Drafting Agent with Quality Review

Reflection and trust patterns

Criteria
Without context
With context

Self-evaluation loop

91%

100%

Iterative revision

30%

20%

Hard iteration limit

100%

100%

No blind output trust

100%

100%

Quality report shows reasoning

70%

80%

Structured action log

90%

100%

Context usage tracked

100%

0%

Scoped task only

100%

100%

Termination condition explicit

100%

100%

Cost control present

62%

62%

Guardrail logged

100%

100%

Without context: $0.4990 · 2m 20s · 23 turns · 217 in / 6,740 out tokens

With context: $0.7813 · 3m 2s · 38 turns · 279 in / 8,547 out tokens

Repository
boisenoise/skills-collections
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.