agent-estimation
by hjw21century
Coding Agents & IDEsBrief
Accurately estimate AI agent work effort using the agent's own operational units (tool-call rounds) instead.
Description
What This Skill Does
Estimates AI agent task effort using tool-call rounds as the primary unit instead of human developer time. Decomposes tasks into modules, applies risk coefficients, and converts to wallclock time only at the final step. Structured output includes a module breakdown table, risk factors, and a time range.
Avoids the systematic overestimation that occurs when agents anchor to human developer timelines from training data, producing estimates grounded in actual agent iteration speed.
When to Use It
- Scoping a coding task before starting an agent session
- Comparing two implementation approaches by round cost
- Setting realistic expectations with a client on agent-assisted work
- Identifying which modules carry the most schedule risk
- Deciding whether a task fits in one session or needs to be split
Example Workflow
User asks: How long will it take to build a custom OAuth2 integration with a third-party API?
- 1
Decompose the task into modules: token exchange flow, refresh logic, API client wrapper, error handling
- 2
Estimate base rounds per module using complexity anchors from the skill table
- 3
Assign risk coefficients based on API documentation quality and platform quirks
- 4
Sum effective rounds across modules and add 10-20% for integration
- 5
Convert total rounds to wallclock time at 3 min/round and output the structured markdown table
Markdown estimation table with per-module round counts, risk levels, integration overhead, and a final wallclock range of 45-60 minutes
Installation
clawhub install hjw21century/agent-estimationRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.