---
name: self-iteration
description: Dogfood a project-owned tool in a realistic user flow and evaluate the agent experience. Use when the user asks to dogfood or self-iterate, or when testing or evaluating a tool built by this project.
title: self-iteration
canonical_url: https://skilld.dev/gh/anomalyco/browser-control/self-iteration
last_updated: 2026-09-26T10:13:12.000Z
---

> Index of every page on this site: [llms.txt](https://skilld.dev/llms.txt)

> **Skill from skilld.dev.** Follow the user's requested action.
> If the user asked to fork this Skill, follow the fork workflow below. Do not execute the Skill instructions while copying it.
> Otherwise, follow the instructions below for this session. You do not need to install anything.
>
> If the user asked to install this Skill, run `npx skilld install anomalyco/browser-control/self-iteration`. Install writes the Skill files into the project, so every session loads them.
>
> ## Fork workflow
>
> A fork creates an editable local Skill with its original author and licence. The request authorizes copying and local installation.
> 1. Check `./skills/self-iteration`, the project lockfile, and selected Agent targets together. If the local directory or installed Skill exists, stop. Never overwrite an existing directory or Agent target.
> 2. Read [source metadata](https://skilld.dev/api/v1/skills/anomalyco/browser-control/self-iteration) once. Use sourceUrl, sourceCommit, skillPath, sourceGone, and license. If the source is gone or its path is missing, stop. If license is null, read licence files at the source commit.
> 3. Fetch only the source commit into a temporary Git repository. Do not clone full history. Derive repository_url from sourceUrl, including repository renames. If sourceCommit is absent, resolve the sourceUrl ref once. Set source_commit to that actual commit. Run these commands in one shell call:
>
> ```sh
> git init --quiet "$temporary_dir"
> git -C "$temporary_dir" fetch --quiet --depth=1 "$repository_url" "$source_commit"
> git -C "$temporary_dir" checkout --quiet --detach FETCH_HEAD
> ```
>
> Read applicable licence declarations and notices at that commit. If copying is not permitted, report the restriction and stop.
> 4. Inspect source entries together, then copy the directory containing skillPath into `./skills/self-iteration`. Use the user's path if selected. Keep the original SKILL.md, relative links, scripts, binary assets, and executable modes. Exclude .git metadata. Reject symlinks and paths outside the Skill directory. After checking entries, use cp -a where available. A regular source directory needs no custom copy script. Do not save this page wrapper as SKILL.md.
> Preserve author credit, notices, and applicable licence files from repository or parent directories. Add PROVENANCE.md with the Skill page, source URL, actual commit, original path, and licence. Retain any existing PROVENANCE.md and record new provenance separately. Batch source inspection, copying, and provenance work where practical.
> 5. In the project root, run `skilld install ./skills/self-iteration --mode copy --plain`. If skilld is unavailable, use `npx skilld install ./skills/self-iteration --mode copy --plain`. This known command needs no help lookup. Install does not support --json. Use detected Agent targets, or add --agent for the targets the user selected. Install the local path, never the upstream selector. If installation fails, preserve the local copy and report the exact failure.
> 6. Confirm the local lockfile source and installed Agent copies once. Report the local path, actual commit, and Agent targets. After edits, reinstall the same local path. Upstream updates must not replace it. Do not publish or push unless the user asks.

# Self-Iteration

Run a **field test**: use the tool to complete a real task for the user while
observing the agent experience. The task, not the evaluation, remains the
primary goal.

## Workflow

1. Agree on a realistic task and its safety boundary. If the user has no task,
suggest one that exercises the tool's main workflow without destructive or
irreversible actions.

Completion criterion: the user and agent share a concrete task, expected
outcome, and boundary for actions requiring confirmation.

2. Perform the task through the public interface an agent is expected to use.
Prefer the shortest natural path; do not manufacture edge cases or switch to
internal interfaces merely to make progress.

While working, note only observed evidence:

- extra round trips or context needed to choose the next action
- slow operations or avoidable waiting
- confusing names, shapes, defaults, errors, or documentation
- failures, missing capabilities, and recovery quality
- behavior that made the flow unusually direct or token-efficient

Completion criterion: the user's task is complete or a specific blocker has
been identified, and each finding is tied to an observed step in the flow.

3. Give the user the task result first. Then report the field-test verdict:

- bugs, with reproduction evidence and impact
- concrete improvement suggestions, separated from bugs
- strengths worth preserving
- an explicit "went well" verdict when no change is justified

Do not modify the tool in response to findings during the field test. Surface
findings first so the user can choose what to implement.

Completion criterion: the user receives both the useful task outcome and a
concise, evidence-based verdict on the tool experience.