---
name: blind-skillroll-evaluation
description: Coordinate an end-to-end independent live evaluation of the current SkillRoll main across a small varied set of external skill repositories.
title: blind-skillroll-evaluation
canonical_url: https://skilld.dev/gh/hagaiw/skillroll/blind-skillroll-evaluation
last_updated: 2026-11-15T17:45:26.000Z
---

> 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.
>
> Supporting files, fetch one when the Skill refers to it: [evals/coordinate-independent-small-repo-trials.eval.md](https://skilld.dev/api/skills-raw/hagaiw/skillroll/blind-skillroll-evaluation/evals/coordinate-independent-small-repo-trials.eval.md), [evals/report-mixed-results-and-product-fix.eval.md](https://skilld.dev/api/skills-raw/hagaiw/skillroll/blind-skillroll-evaluation/evals/report-mixed-results-and-product-fix.eval.md), [references/context.md](https://skilld.dev/api/skills-raw/hagaiw/skillroll/blind-skillroll-evaluation/references/context.md).
>
> If the user asked to install this Skill, run `npx skilld install hagaiw/skillroll/blind-skillroll-evaluation`. 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/blind-skillroll-evaluation`, 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/hagaiw/skillroll/blind-skillroll-evaluation) 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/blind-skillroll-evaluation`. 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/blind-skillroll-evaluation --mode copy --plain`. If skilld is unavailable, use `npx skilld install ./skills/blind-skillroll-evaluation --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.

# Blind SkillRoll evaluation

Use the installed `flow-runner` to compose `select-blind-targets`,
`prepare-blind-trial`, `author-blind-evals`, and `run-blind-evals`. Execute the
campaign rather than returning only a plan.

With a generic action wrapper, request exactly the `Skill` action with arguments
`{"name":"flow-runner"}`. Do not request the four component names directly;
the installed runner coordinates them and returns their artifacts.

Freeze selection before outcomes, use exact revisions in isolated temporary
workspaces, and give each target's authoring worker only its neutral manifest
and permitted skill bundle. When fresh agents or contexts are available, use
one per target and do not seed them with previous reports or suspected bugs.
Run the frozen cases with authorized live inference, then reconcile evidence in
the coordinating context.

Report setup defects, case-authoring defects, target behavior, product defects,
and provider errors separately. A product defect may justify a small regression
fix in SkillRoll when the user authorized changes to this repository; it does
not erase the external run that exposed it.

Lead the results section with the observed count for each outcome, including
errors; then explain case-authoring issues, target behavior, product evidence,
and provider problems separately. State whether a control or other isolating
evidence supports a product conclusion. Without it, say that the product cause
is unresolved rather than attributing target outcomes to SkillRoll.

Read [campaign context](https://skilld.dev/api/skills-raw/hagaiw/skillroll/blind-skillroll-evaluation/references/context.md) for phase handoffs and the final
report.

Keep generated CI advisory. Report whether CI was generated and its observed
status, along with whether repository commands, commits, pushes, or publishing
ran.
