claude-feature-adoption

name	claude-feature-adoption
description	Review what shipped in Claude Code since the last adoption analysis and file an umbrella issue (☂️) + one subissue per adoptable feature, each grounded in how Sybra actually drives Claude headless. Use when asked to "check what's new in Claude Code", "adopt new Claude features", or to refresh the Claude Code adoption tracker. For the Codex CLI, use codex-feature-adoption instead.
argument-hint	[--from <version>] [--dry-run]
user-invocable	true
allowed-tools	["Agent","Read","Grep","Glob","Bash","WebFetch"]

Claude Code Feature Adoption

Find what shipped in Claude Code since Sybra last did an adoption pass, decide what's worth adopting (grounded in how Sybra actually invokes the claude CLI), and file an umbrella issue with one subissue per item.

This is the Claude variant. For the OpenAI Codex CLI, use codex-feature-adoption — same workflow, different changelog source and grounding points.

Repo: Automaat/sybra (public personal project). Everything this skill writes is about Claude Code public features — safe to file. Never paste work-repo content (see the Work-Data Confidentiality rule in the root CLAUDE.md); only reference Sybra's own code.

Phase 1 — Establish the baseline (what's already covered)

1. Find the last umbrella. List prior adoption trackers, newest first:

   gh issue list --repo Automaat/sybra --state all --search '"Adopt Claude Code" in:title' \
     --json number,title,state,createdAt --limit 20
   

   The most recent ☂️ Adopt Claude Code vX-Y features gives the floor = its upper version Y. If none exists, ask the user for a starting version or default to ~1 month back.

2. Read the last umbrella + its subissues so you don't re-file already-adopted items:

   gh issue view <last-umbrella> --repo Automaat/sybra --json title,body
   gh api graphql -f query='{ repository(owner:"Automaat",name:"sybra"){ issue(number:<last-umbrella>){ subIssues(first:50){ nodes{ number title state } } } } }' \
     -H "GraphQL-Features: sub_issues"
   

   Closed subissues = already adopted (skip). Open subissues = carryovers to list in the new umbrella, noting any the new changelog reinforces or unblocks.

3. Current installed version = the ceiling:

   claude --version
   

   New range = (floor + 1) → installed. Override the floor with --from <version> if passed.

Phase 2 — Read the changelog for the new range only

• Primary source: the local cache ~/.claude/cache/changelog.md. Version headers are ## X.Y.Z, newest first. Find the line for the floor version and read everything above it up to the installed version. It's large — read in slices with offset/limit; never dump the whole file.

  grep -n '^## ' ~/.claude/cache/changelog.md | head -80   # locate the range line numbers
  

• Fallback (cache missing/stale): WebFetch the official changelog — https://raw.githubusercontent.com/anthropics/claude-code/main/CHANGELOG.md (or the @anthropic-ai/claude-code npm page).

Phase 3 — Ground the analysis in Sybra's code

A changelog entry only matters if it touches how Sybra invokes or manages claude. Re-verify against the current tree every run (the map drifts) — spawn Explore agents in parallel, or read directly. Known integration points (verify, don't trust):

Concern	Where to look
Headless claude -p flags	internal/agent/runner_headless_invocation.go (the non-codex branch: -p, --output-format stream-json --verbose, --resume, --allowedTools / --dangerously-skip-permissions, --model)
Subprocess env vars	same file — BASH_*_TIMEOUT_MS, CLAUDE_CODE_FORK_SUBAGENT; check what's not set (fallback model, retry watchdog, effort, session id, AI_AGENT)
NDJSON stream parsing	internal/agent/runner_headless.go, internal/agent/stream.go — which event types (system/init, assistant, user, result) and fields (plugin_errors) are handled
Resume / session id	manager_run.go, runner_headless.go (session id captured from init)
Approval hook	internal/agent/runner_convo.go + internal/agent/approval_server.go (HTTP PreToolUse via --settings)
Scrub / confidentiality	internal/scrub, internal/sybra/app_human_review.go, monitor_sink.go
Model list	internal/sybra/svc_info.go, manager_run.go normalizeModel (Claude side is hardcoded aliases)
Reviews	internal/sybra/app_reviews_inbound.go (/staff-code-review, /fix-review-auto)
Server / unattended	root CLAUDE.md "Server Deployment" — unattended headless agents on the synapse LXC

For each candidate feature, state exactly where it plugs in (file:line) and the effort. Drop anything UI-only, terminal-rendering, Windows-specific, or already adopted.

Phase 4 — Bucket the findings

Mirror the prior umbrellas (#685, #1019):

• High value — clean adopts that improve server resilience, reliability, or cost at low effort (e.g. --fallback-model, CLAUDE_CODE_RETRY_WATCHDOG, --json-schema structured output).
• Medium — worthwhile but more effort or a behavior change (e.g. auto mode vs --dangerously-skip-permissions, --effort per role, disabling bundled skills, model-picker additions).
• Smaller — confidentiality/security/cleanup (e.g. attribution.sessionUrl, sandbox.credentials, free bug-fix-on-bump).
• Parallel groups — Group A = changes that all touch the same file (e.g. runner_headless_invocation.go) so they serialize; Group B = independent, parallelizable.
• Carryovers — still-open subissues from prior umbrellas; flag which the new changelog reinforces/unblocks.
• Strategic watch — large/convergent items, not drop-ins (e.g. native claude agents/--bg, OTEL signals for the monitor).
• Skipped — features evaluated and rejected, with the one-line reason.

Present this to the user before filing. If invoked --dry-run, stop here.

Phase 5 — File the umbrella + subissues

1. Umbrella (title prefixed with ☂️, per the global git rules):

   gh issue create --repo Automaat/sybra \
     --title "☂️ Adopt Claude Code vX-Y features" \
     --body-file <umbrella.md>
   

   Body = the buckets above + parallel groups + carryovers (with #refs) + a ## Source line naming the changelog range. Write bodies to files in the scratchpad; never use -F/heredoc-to---body for long text.

2. One subissue per adoptable item. Each body: the bucket + parallel group, the Sybra grounding (file refs), concrete steps, and a Source: Claude Code CHANGELOG vX.Y.Z line. Create, then link to the umbrella:

   gh-add-subissue <umbrella#> <child#> Automaat/sybra
   

   (gh-add-subissue is in ~/.local/bin; it needs the GraphQL-Features: sub_issues header, which it sets.)

3. Verify the links landed — gh-add-subissue occasionally drops one:

   gh api graphql -f query='{ repository(owner:"Automaat",name:"sybra"){ issue(number:<umbrella#>){ subIssues(first:50){ totalCount } } } }' \
     -H "GraphQL-Features: sub_issues"
   

   If totalCount < number created, re-run gh-add-subissue for the missing child.

4. Report the umbrella URL and the subissue list, and recommend the highest-leverage / lowest-effort starting pair.

Conventions

• Commit/issue titles: conventional-commit style where applicable; umbrella titles lead with ☂️.
• No work-repo identifiers anywhere. No AI attribution. No #-ref noise that title hooks reject.
• Keep subissue bodies terse and grounded — a generic body ("adopt feature X") is a failure; name the file it lands in.