---
name: adobe-performance-tuning
description: >-
  Improve Adobe integration latency and throughput from measurements while preserving correctness, policy, and spend boundaries. Use when the task requires adobe measured performance tuning. Trigger with "tune Adobe performance", "reduce Firefly latency", or "optimize PDF jobs".
allowed-tools: Read,Glob,Grep,Write,Edit
argument-hint: "<service-operation> <measurement-window> <objective>"
version: 1.8.0
license: MIT
author: Jeremy Longshore <jeremy@intentsolutions.io>
tags: [ saas, adobe, performance ]
model: inherit
effort: high
compatibility: "Designed for Claude Code; live Adobe actions require network access, appropriate entitlement and authentication, and explicit approval"
title: adobe-performance-tuning
canonical_url: https://skilld.dev/gh/jeremylongshore/tons-of-skills-marketplace/adobe-performance-tuning
last_updated: 2026-10-07T06:06:28.000Z
---

> 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: [references/official-docs.md](https://skilld.dev/api/skills-raw/jeremylongshore/tons-of-skills-marketplace/adobe-performance-tuning/references/official-docs.md).
>
> If the user asked to install this Skill, run `npx skilld install jeremylongshore/tons-of-skills-marketplace/adobe-performance-tuning`. 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/adobe-performance-tuning`, 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/jeremylongshore/tons-of-skills-marketplace/adobe-performance-tuning) 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/adobe-performance-tuning`. 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/adobe-performance-tuning --mode copy --plain`. If skilld is unavailable, use `npx skilld install ./skills/adobe-performance-tuning --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.

# Adobe Measured Performance Tuning

## Overview

Improve Adobe integration latency and throughput from measurements while preserving correctness, policy, and spend boundaries. This workflow produces a reviewable artifact and evidence before any live side effect.

## Prerequisites

- Current first-party Adobe documentation for every selected service, API version, auth flow, limit, and lifecycle.
- Named product, identity, security, data, budget, release, and operations owners appropriate to the scope.
- Synthetic or approved non-production fixtures with secret and content canaries.

## Current Contract

No universal Adobe latency table is a production contract. Measure token reuse, queue wait, upload/download, submission, status polling, vendor processing, validation, and downstream work separately by service and operation. Recheck the dated evidence map before relying on mutable product behavior.

## Authentication

Use content-free metrics and aliases. Token reuse must honor actual expiry response and revocation; cached signed URLs and customer outputs are not performance caches by default.

## Instructions

1. Define objective, workload, input class, concurrency, cost ceiling, completeness, and measurement window.
2. Instrument queue, transport, vendor job, polling, storage transfer, validation, retry, and downstream spans.
3. Measure percentiles, 429s, failures, bytes, transactions/generations, and artifact correctness from a synthetic canary.
4. Test token reuse, connection reuse, right-sized inputs, bounded concurrency, adaptive polling, and safe deduplication independently.
5. Canary one change within current documented constraints and compare correctness plus spend, not latency alone.
6. Retain verified gains with rollback thresholds and remove instrumentation that captures sensitive content.

## Tool Discipline

Use Read, Glob, and Grep to inspect current documentation, configuration, code, fixtures, and evidence. Use Write and Edit only for approved repository artifacts. Skill invocation alone does not authorize network access, credentials, Adobe content, consent, uploads, generation, spend, deployment, registration changes, replay, cancellation, or deletion.

## Approval Boundaries

Workload, data, and budget owners approve load or generation tests. Do not increase credentials or identities to manufacture throughput.

## Error Handling

- Delete invented benchmark tables.
- Do not parallelize past queue, spend, or service evidence.
- Roll back if output correctness, content provenance, throttling, or cost worsens.

## Output

Return baseline spans, bottleneck attribution, experiments, before/after percentiles, correctness/spend proof, selected change, and rollback. Mark assumptions, observed environment behavior, owners, evidence dates, and unresolved gaps explicitly.

## Examples

- Compare fixed polling with bounded response-led adaptive polling.
- Prove token reuse stops on actual expiry or revocation.

## Validation

Exercise and record expected and observed results for:

- queue wait
- upload bottleneck
- 429
- unknown status
- cache staleness
- rollback

## Resources

- [Current first-party evidence map](https://skilld.dev/api/skills-raw/jeremylongshore/tons-of-skills-marketplace/adobe-performance-tuning/references/official-docs.md) — recheck dated Adobe sources before execution.
- Treat observed tenant or product behavior as environment-specific evidence, never a universal Adobe guarantee.
