ai-agents-architect

Expert in designing and building autonomous AI agents. Masters tool use, memory systems, planning strategies, and multi-agent orchestration. Use when: build agent, AI agent, autonomous agent, tool ...

Install with Tessl CLI

npx tessl i github:sickn33/antigravity-awesome-skills --skill ai-agents-architect

What are skills?

1.13x

Quality

51%

Does it follow best practices?

Impact

82%

1.13x

Average score across 3 eval scenarios

Optimize this skill with Tessl

npx tessl skill review --optimize ./skills/ai-agents-architect/SKILL.md

SKILL.md

Review

Evals

Evaluation results

87%

23%

Research Assistant Agent

ReAct loop with iteration limits and error handling

Criteria

Without context

With context

Thought step present

60%

100%

Action step present

87%

100%

Observation step present

87%

100%

Max iteration limit enforced

60%

100%

Graceful stop at limit

90%

100%

Tool error handling

25%

Error surfaced to agent

20%

60%

Trace log written

91%

100%

Robust output parsing

100%

No unlimited autonomy

57%

100%

Without context: $0.5558 · 2m 34s · 24 turns · 28 in / 8,725 out tokens

With context: $0.8673 · 3m 46s · 38 turns · 42 in / 12,399 out tokens

62%

-2%

Developer Productivity Agent Toolkit

Tool registry with complete specs and selective tool loading

Criteria

Without context

With context

Tool Registry class/module

100%

Schema in tool registration

Examples in tool specs

Complete tool descriptions

20%

Tool selector implemented

100%

Avoids tool overload

100%

Lazy loading mechanism

100%

Usage tracking

100%

usage_report.json produced

100%

Tool errors surfaced

Without context: $1.0935 · 3m 57s · 44 turns · 53 in / 15,081 out tokens

With context: $0.8344 · 3m 44s · 29 turns · 282 in / 14,291 out tokens

99%

11%

Automated Report Generation Pipeline

Plan-and-execute with durable workflow state

Criteria

Without context

With context

Distinct planning phase

100%

plan.json written

100%

Execution follows plan

50%

100%

Replanning on failure

40%

100%

Checkpoint file written per step

100%

Resumption from checkpoint

100%

Execution log traces steps

90%

100%

Selective memory

87%

Separate planner component

100%

report.md produced

100%

Agent internals traceable

100%

Without context: $0.4036 · 1m 45s · 15 turns · 17 in / 7,251 out tokens

With context: $0.7864 · 2m 55s · 33 turns · 69 in / 11,709 out tokens

Evaluated: 4 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Research Assistant Agent Developer Productivity Agent Toolkit Automated Report Generation Pipeline

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.