CtrlK
BlogDocsLog inGet started
Tessl Logo

sharaf/migrate-to-tessl

Use when migrating, restructuring, publishing, or auditing an existing Claude skill into a Tessl tile; converting flat .md files or SKILL.md bundles; fixing Tessl Quality, Impact, Uplift, frontmatter, metadata, tile.json summary, README, markdown reference links, registry-vs-local Quality gaps, artifact anchors, auto-eval wait discipline, or pushing tile scores from 88-99% to 100%.

100

2.56x
Quality

100%

Does it follow best practices?

Impact

100%

2.56x

Average score across 3 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

Overview
Quality
Evals
Security
Files

SKILL.mdskills/migrate-to-tessl/

name:
migrate-to-tessl
description:
Use when migrating, restructuring, publishing, or auditing an existing Claude skill into a Tessl tile; converting flat .md files or SKILL.md bundles; fixing Tessl Quality, Impact, Uplift, frontmatter, metadata, tile.json summary, README, markdown reference links, registry-vs-local Quality gaps, artifact anchors, auto-eval wait discipline, or pushing tile scores from 88-99% to 100%.
metadata:
{"version":"0.2.10","source_session":"agent-prompt-engineer registry score verification, 2026-05-22"}

Migrate to Tessl

Purpose

Turn an existing Claude skill into a published Tessl tile. Target registry Quality 100% and with-context/Impact 100% when scenario rubrics are controllable.

Workflow

#PhaseHands off toReference
1Preflight + inventory the source skill(preflight only)phase-1-preflight.md
2Scaffold the tile structuretessl__tile-creatorphase-2-scaffold-tile.md
3Migrate content with progressive disclosure(structural fix)phase-3-migrate-content.md
4Fix frontmatter, sync tile.json, add README(structural fix)phase-4-frontmatter-and-readme.md
5Initial publish + registry Quality 100 gate(tessl CLI + API)phase-5-initial-publish.md
6Evals + optimization looptessl__setup-skill-performance then tessl__optimize-skill-performance-and-instructionsphase-6-evals-and-optimize.md
7Multi-model variance check + final publishtessl__compare-skill-model-performancephase-7-multi-model-verify.md

Always Check First

  • prerequisite-skills.md before Phase 1. Stop if tessl, login, or required Tessl skills are missing.
  • lessons-learned.md before editing. It covers cached evals, registry-vs-local Quality gaps, artifact anchors, auto-eval waiting, and model variance.

Command Anchors

Use these directly when the phase calls for them:

tessl --version
tessl whoami
tessl tile lint <tile-path>
tessl tile publish --dry-run <tile-path>
tessl tile info <workspace>/<tile-name>[@<version>]
tessl eval view <eval-run-id>
tessl search --json <tile-name>

Critical Inline Path

Use this minimum path even before opening details:

  1. Inventory source layout; choose tile root by repo convention, not a fixed tessl/ folder.
  2. Put execution guidance in SKILL.md; move depth to linked references/.
  3. Normalize frontmatter, sync tile.json.summary, and add root README.md.
  4. Run lint, local skill review, and publish dry-run before first publish.
  5. Publish, wait for moderation, then fetch /skill-reviews; fix every dimension below 3/3 before eval work.
  6. Wait for auto-eval completion; use tessl search --json to confirm aggregate, Quality, Impact, and eval average are all 1 for total-score targets. Report Security separately.

Source Triage

  • Flat .md file: all 7 phases.
  • Folder with SKILL.md + references/: preserve references; see Phase 3.
  • Existing tile (tile.json present): skip scaffold/migration and jump to Phase 5 or 6.
  • A .skill archive (zipped, legacy): unzip first, then treat as flat .md.

Done when

  • Published at <workspace>/<tile-name>@<version> with Passed Moderation ✔
  • Registry Quality is 100% before eval optimization starts
  • Exact registry Quality report was fetched from /skill-reviews; every judge dimension is 3/3 and validation passed
  • tessl search --json reports aggregate, Quality, Impact, and eval average at 1.0 for total-score targets
  • All scenarios at 100% with-context on at least one model, ≥ 95% on others
  • Uplift/Impact recorded, with multiplier interpreted against the observed baseline
  • Activation fires on every scenario (no routing gaps)
  • Local diff is committable

Stop conditions

Stop and report when preflight fails, source parsing fails, Tessl project setup fails, scenario generation hangs >30 min on a second attempt, or three consecutive eval runs stay below 80% with-context.

Otherwise, proceed phase by phase. Each phase reference file has explicit success/failure criteria.

skills

migrate-to-tessl

SKILL.md

README.md

tile.json