Fastly Compute Publish Ignores Draft Versions

What it does

When your agent uses it

• You ran fastly service backend create --autoclone (or any dashboard/CLI change that creates a draft version) but did not activate that draft
• Compute code returns backend-not-found errors
• Misroutes requests
• Youre surprised to see your draft version number skipped over in the version chain

Example prompts

• “backend exists on a draft version but isn”
• “re surprised to see your draft version number”
• “/fastly-compute-publish-ignores-draft-versions”

Workflow steps

4 steps, taken from the first numbered list in SKILL.md.

1. 1Do all non-code changes through fastly compute publish first, then WASM changes.
2. 2Always activate your draft before running fastly compute publish. If you ran
3. 3Treat every draft version as an unshipped change. Before any fastly compute publish,
4. 4Use --verbose on fastly compute publish. The verbose output shows which version

What it can do on your machine

Read from SKILL.md and the folder at commit c3d6f7e. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  No scripts in the folder and no shell commands in SKILL.md (its code samples are bash).

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Fastly Compute Publish Ignores Draft Versions loads about 2.3k tokens when it runs. Until then it costs about 205 tokens; SKILL.md has 851 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~205

When it runs · the whole SKILL.md, loaded when a task matches

~2.3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from divinevideo/divine-mobile at commit c3d6f7e, republished under its MPL-2.0 licence (© divinevideo). 851 words, ~2,292 tokens.

name

fastly-compute-publish-ignores-draft-versions

description

Fix "backend exists on a draft version but isn't reachable from the new Compute deploy" bugs after `fastly compute publish`. Use when: (1) You ran `fastly service backend create --autoclone` (or any dashboard/CLI change that creates a draft version) but did not activate that draft, (2) You then ran `fastly compute publish` which appeared successful but your backend / ACL / header / dictionary change is missing from the active version, (3) Compute code returns backend-not-found errors or misroutes requests, (4) You're surprised to see your draft version number "skipped over" in the version chain. Root cause: `fastly compute publish` clones from the currently-active version, not from the latest draft, so dashboard changes made in an un-activated draft get stranded.

author

Claude Code

version

1.0.0

date

2026-04-05

t0  Active = v250
t1  $ fastly service backend create --version latest --autoclone \
       --name cloud_run_transcoder --address ...
    → creates v251 (clone of v250) + adds backend
    → v251 is DRAFT, not active
    → "latest" in this context means "highest version number" (v251), not "active version"

t2  $ fastly compute publish --comment "ship fix"
    → creates v252 (clone of v250, NOT v251) + new WASM
    → stages and activates v253 (clone of v252) + activates
    → v251's backend change was never carried forward
    → active = v253, which does NOT contain cloud_run_transcoder

t3  Compute code at v253 calls send_async("cloud_run_transcoder")
    → backend name does not exist on v253
    → silent failure, fire-and-forget hides it

# Find all versions and which is active
fastly service version list --service-id YOUR_SERVICE_ID

# For each recent version, list backends (or whatever resource you added)
for v in 250 251 252 253 254; do
  echo "=== v$v ==="
  fastly service backend list --service-id YOUR_SERVICE_ID --version $v | grep MY_BACKEND_NAME
done

# Creates a new draft cloned from the active version, with the backend added.
# Activate the draft atomically afterward.
fastly service backend create --service-id YOUR_SERVICE_ID --version latest --autoclone \
  --name cloud_run_transcoder \
  --address divine-transcoder-149672065768.us-central1.run.app \
  --port 443 --use-ssl \
  --ssl-sni-hostname divine-transcoder-149672065768.us-central1.run.app \
  --override-host divine-transcoder-149672065768.us-central1.run.app

# Confirm the draft has the backend
fastly service backend list --service-id YOUR_SERVICE_ID --version LATEST_DRAFT \
  | grep cloud_run_transcoder

# Activate
fastly service version activate --service-id YOUR_SERVICE_ID --version LATEST_DRAFT

# Purge cache so any poisoned 404/502 responses drop
fastly purge --all --service-id YOUR_SERVICE_ID

# Active version has the backend
fastly service backend list --service-id YOUR_SERVICE_ID --version $(fastly service version list --service-id YOUR_SERVICE_ID | awk '$3=="true"{print $1}') | grep MY_BACKEND_NAME

# End-to-end: hit a Compute endpoint that exercises the backend and check logs
# for successful routing (no "backend does not exist" errors)
fastly log tail --service-id YOUR_SERVICE_ID | grep -iE "backend|MY_BACKEND"

© divinevideo, MPL-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .agents/skills/fastly-compute-publish-ignores-draft-versions of divinevideo/divine-mobile.

Open the folder on GitHubat commit c3d6f7e