The post-session summary (shown after every SSH session ends) now:
- Displays the server name when available, so users can find it in their
cloud dashboard (e.g., "Your server 'spawn-claude-abc' is still running")
- Adds explicit billing reminder ("Remember to delete it to avoid charges")
- Uses green (log_info) for reconnect instructions instead of yellow
(log_warn), since reconnect info is helpful guidance, not a warning
No changes to individual cloud scripts needed -- all scripts already set
SERVER_NAME before calling interactive_session.
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix(ci): propagate mock test exit code and fix broken pipe in summary
The test workflow had three issues:
- mock.sh exit code was swallowed by tee (no pipefail), so the check
always passed even with 165 failures
- grep|head pipe caused "write error: Broken pipe" in post summary
- Summary was noisy with 100+ individual result lines
Now uses PIPESTATUS[0] to capture the real exit code, shows a clean
results line plus collapsible failures list, and fails the check when
tests fail.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(ci): report test results without blocking PRs
Pre-existing failures (165) shouldn't block unrelated PRs. The summary
still shows pass/fail counts and a collapsible failures list so the bot
can see the results.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* perf(ci): increase QA cycle frequency from daily to every 4 hours
Daily runs meant breakage could go undetected for up to 24 hours.
Every 4 hours gives 6 runs/day (00:00, 04:00, 08:00, 12:00, 16:00,
20:00 UTC) with a max 4-hour feedback loop.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(ci): add missing Check results step to fail on test errors
Addresses review feedback:
- The exit code was captured via PIPESTATUS[0] into GITHUB_OUTPUT but
no subsequent step consumed it, so the workflow always passed even
when tests failed. Added a "Check results" step that reads the
captured exit code and fails the job accordingly.
- Reverted QA cron schedule change (every 4 hours back to daily at
06:00 UTC) as it was unrelated to the test exit code fix and should
be proposed separately if desired.
Agent: pr-maintainer
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Cover the _show_post_session_summary function and updated
ssh_interactive_session integration from PR #1037. Tests verify:
- Summary warns user their server is still running with IP
- Dashboard URL shown when SPAWN_DASHBOARD_URL is set
- Generic message when no dashboard URL is available
- Reconnect command uses correct SSH_USER and IP
- SSH exit code preserved through the summary display
- All 25 SSH-based cloud providers set SPAWN_DASHBOARD_URL
- SPAWN_DASHBOARD_URL uses HTTPS and is defined before usage
- Detects custom interactive_session implementations missing summary
(alibabacloud flagged as known gap)
Agent: test-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Extract `_aliyun_json_list_first` helper for flat JSON lists (unlike
`_aliyun_json_field` which handles lists of dicts)
- Extract `_aliyun_extract_instance_id` to replace inline Python parser
- Extract `_ensure_network_infrastructure` to consolidate VPC/vSwitch/SG setup
- Use `_log_diagnostic` for structured error reporting (consistent with
patterns in shared/common.sh)
Reduces create_server from 86 to 69 lines and eliminates inline Python.
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
5 improvements to the QA cycle:
1. Fix agents now get structured failure context — categorized failures
(exit_code, missing_api_call, missing_env, no_fixture) instead of
raw 500-line test output, plus a passing agent for comparison
2. Fix agent changes are verified before committing — re-runs mock tests
after the agent finishes and only commits if results actually improved,
discarding bad fixes that would create noise PRs
3. Test results now include failure categories — mock.sh records
cloud/agent:fail:reason instead of just cloud/agent:fail, enabling
smarter failure routing
4. Mock curl logs NO_FIXTURE warnings when no fixture matches a GET
request, surfacing false-confidence gaps where tests pass with
synthetic fallback data
5. Phase 3 (code fix) failures now escalate to GitHub issues after 3
consecutive cycles, matching the Phase 1 escalation pattern
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
After an interactive SSH session ends, users are now shown:
- A warning that their server is still running (and may incur charges)
- A link to the cloud provider's dashboard to manage/delete it
- The SSH command to reconnect
This prevents users from unknowingly leaving servers running after
exiting their agent session. Covers all 25 SSH-based cloud providers.
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Unquoted `<< EOF` heredocs in nanoclaw .env file creation cause shell
expansion of the API key value. If an API key contains `$`, backticks,
or `\`, the value is silently corrupted or could trigger command
execution. Replace with `printf '%s'` which safely writes the value
without interpretation.
Also fix unquoted variable expansion in upload_config_file's mv command
and the github-codespaces/openclaw.sh config heredoc.
Fixes 34 scripts across all cloud providers.
Agent: security-auditor
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add validate_branch_name() and validate_cloud_name() to qa-cycle.sh to
prevent command injection via unvalidated strings passed to git/gh
commands. Cloud names parsed from test/record.sh output via sed were
used directly in branch names, git push, git worktree, and gh pr create
commands without validation.
Fixes#1028
Agent: security-auditor
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Validates CloudSigma's unique architecture: region-based API URLs,
HTTP Basic Auth (email + password), drive cloning workflow, python3
JSON construction, SSRF-preventing region validation, and SSH with
'cloudsigma' user. Covers lib/common.sh API surface, all 8 agent
scripts, manifest consistency, and test infrastructure (mock.sh +
record.sh).
Agent: test-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When spawn scripts fail or are interrupted, error messages now include
the cloud provider's actual dashboard URL instead of generic "check your
cloud provider dashboard" text. This helps users quickly navigate to
their provider to check server status, clean up orphaned resources, or
debug provisioning failures.
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The interactive flow (bare `spawn`) was missing the preflight credential
warning that the direct `spawn <agent> <cloud>` path already had. Users
who picked an agent and cloud interactively would not be warned about
missing credentials, leading to confusing failures from the cloud
provider script. Now both paths warn about missing credentials before
launching.
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract _aliyun_load_or_prompt_credentials and _aliyun_configure_cli from
the 68-line ensure_aliyun_credentials function, reducing it to 16 lines.
Extract _aliyun_instance_public_ip to replace inline Python in
_wait_for_aliyun_instance, making IP extraction reusable and consistent
with the existing _aliyun_json_field helper pattern.
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add path validation to --prompt-file to block reading sensitive files
(SSH keys, cloud credentials, .env files, etc.) whose contents would be
sent to remote agents. Also adds file size validation (1MB limit) and
stat-based file type checking.
Fixes#991
Agent: security-auditor
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When users run `spawn` interactively, the agent picker now shows how many
clouds each agent supports and how many have credentials ready. This helps
users quickly identify which agents they can deploy immediately.
Before: "Claude Code AI coding assistant"
After: "Claude Code 2 clouds, 1 ready"
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
When running `spawn claude` or `spawn hetzner`, the "Loading manifest..."
spinner appeared twice: once in showInfoOrError() and again in
cmdAgentInfo/cmdCloudInfo via validateAndGetEntity(). Pass the
pre-loaded manifest to avoid the redundant load and spinner flash.
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract validation, error handling, and response parsing from
create_server into dedicated helpers following the pattern from PR #1016.
Latitude helpers: _latitude_validate_inputs, _latitude_check_create_error,
_latitude_extract_server_id
Contabo helpers: _contabo_validate_inputs, _contabo_check_create_error,
_contabo_extract_instance_id
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add validate_resource_name check for ALIYUN_IMAGE_ID env var in
alibabacloud create_server, consistent with other providers (Contabo,
Webdock) that validate user-controllable image identifiers
- Move HOSTKEY location validation before _pick_instance_preset call,
which uses the location in an API request — validates input before
use rather than after
Agent: security-auditor
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract _ensure_vpc, _ensure_vswitch, _aliyun_json_field, and
_aliyun_json_top_field from the 182-line create_server function.
This reduces create_server to 85 lines and eliminates repeated
inline Python JSON parsing across multiple functions.
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add regex validation (^[a-z0-9][a-z0-9._-]{0,63}$) to invalidate_cloud_key()
in shared/key-request.sh to prevent path traversal attacks that could delete
arbitrary files via crafted provider names (e.g., ../../etc/important)
- Improve validKeyVal() in key-server.ts to block control characters
(U+0000-U+001F, U+007F-U+009F) and enforce a 4096-byte max length on
API key values, preventing injection of null bytes, newlines, and
excessively long values
Agent: security-auditor
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
webdock/lib/common.sh:
- Extract _webdock_get_public_key_ids() for SSH key ID fetching
- Extract _webdock_validate_inputs() for input validation
- Extract _webdock_handle_create_response() for response parsing and error reporting
- create_server reduced from 53 to 24 lines
koyeb/lib/common.sh:
- Extract _koyeb_detect_os() for OS detection
- Extract _koyeb_detect_arch() for architecture detection
- Extract _koyeb_install_cli() for download and PATH setup
- ensure_koyeb_cli reduced from 51 to 13 lines
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When a spawn script is killed by a signal (SIGKILL, SIGTERM, SIGHUP, etc.),
Node.js returns exit code null. Previously this produced the confusing message
"Script exited with code null". Now detects the actual signal and shows
signal-specific guidance: OOM suggestions for SIGKILL, terminal reconnection
tips for SIGHUP, spot instance warnings for SIGTERM.
Fixes#1011
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Webdock was added in PR #1001 with zero dedicated test coverage.
This adds comprehensive tests validating:
- lib/common.sh API surface (required + provider-specific functions)
- API base URL and constants
- Credential handling (ensure_api_token_with_provider pattern)
- SSH key management (json_escape for injection prevention)
- Server lifecycle (generic_cloud_api, generic_wait_for_instance)
- SSH delegation pattern (ssh_run_server, ssh_upload_file, etc.)
- Security conventions (no echo -e, no set -u, validate_resource_name)
- Agent script patterns (claude, aider, cline)
- Manifest consistency (type, auth, exec_method, defaults)
- Test infrastructure coverage (mock.sh and record.sh entries)
Agent: test-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Change 'auth: TOKEN' to 'needs TOKEN' with yellow highlight in spawn clouds
- Always show legend footer explaining ready/needs indicators
- Add --clear hint to spawn list footer
- Show --version/-v and --help/-h aliases in help text
- Document SPAWN_UNICODE=1 env var in help
- Include HTTP status code in update fetch errors
- Bump version to next patch
Fixes#1010
Agent: issue-fixer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
Tests cover the recently decomposed helper functions from PR #976
(cmdAgentInfo, generic_wait_for_instance) to ensure the refactored
helpers maintain correct behavior.
Agent: test-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Haiku 4.5 <noreply@anthropic.com>
Extract _get_multi_cred_spec, _load_multi_config_from_file, and
_save_multi_config_to_file helpers to eliminate duplicated per-cloud
config blocks in try_load_config, save_config, has_credentials,
prompt_credentials, and list_clouds.
The cloud-to-credential mapping (OVH, UpCloud, Kamatera, AtlanticNet,
CloudSigma) is now defined once in _get_multi_cred_spec and consumed
by all five functions, making it trivial to add new multi-credential
clouds.
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract three helpers from the 82-line, 14-conditional function:
- _parse_cloud_auths: extract cloud auth specs from manifest.json
- _try_load_env_var: load a single env var from env or config file
- _load_cloud_credentials: load all env vars for one cloud provider
The main function is now a 36-line orchestrator with clear flow:
validate prerequisites -> parse manifest -> iterate clouds -> summarize.
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Haiku 4.5 <noreply@anthropic.com>
Cover printAgentQuickStart (commands.ts) which has zero test coverage:
- Single-auth and multi-auth cloud credential display
- URL hint placement (only on first auth var)
- All/partial/no credentials detection ("ready to go" vs export lines)
- No-auth cloud (auth="none") handling
- Agent info header, install line, available clouds listing
- Credential prioritization in cloud ordering
- Grouped cloud type display and credential indicators
- Pure logic replica tests for quick-start computation
Agent: test-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Implements Webdock cloud provider with full API integration:
- webdock/lib/common.sh with REST API primitives
- claude.sh, cline.sh, aider.sh agent scripts
- Test coverage in test/record.sh and test/mock.sh
- manifest.json updated with cloud entry and matrix
- README.md with usage documentation
Webdock offers affordable European VPS (€2.15/month starting) with
full REST API, SSH access, and developer-friendly features.
Agent: cloud-scout-1
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
- Show list-specific flags (-a, -c, --clear) in unknown flag error
- Add specific error for empty prompt files instead of generic validation
- Document SPAWN_UNICODE=1 env var in help text and troubleshooting
- Show filter/clear hints in interactive list picker
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Implements CloudSigma matrix entries for openclaw, nanoclaw, interpreter, continue, gemini, and codex. All scripts follow the standard CloudSigma pattern with OpenRouter API key injection.
Agent: gap-filler
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
- fly/lib/common.sh: Replace 23-line get_server_name() that duplicated
env-var-check, prompt, and validation logic with a one-line call to the
shared get_validated_server_name helper, matching all other cloud providers.
- oracle/lib/common.sh: Break _setup_vcn_networking (48 lines, 3 distinct
responsibilities) into focused helpers:
- _create_internet_gateway: creates the IGW resource
- _add_default_route: configures the route table
- _add_ssh_security_rules: opens SSH port in the security list
The orchestrator _setup_vcn_networking now delegates to these three helpers.
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Ensures closing a PR also deletes its remote branch, consistent with
how gh pr merge already uses --delete-branch. Removes redundant manual
git push origin --delete calls that were previously needed.
Fixes#942
Agent: pr-maintainer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
SLACK_WEBHOOK was embedded directly in heredocs at three locations,
allowing potential command injection if the env var contained shell
metacharacters. Added early validation requiring the URL to match
the expected Slack webhook format (https://hooks.slack.com/...).
Also stopped leaking the full webhook URL into prompt text.
Fixes#992
Agent: security-auditor
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Validates that test/mock.sh and test/record.sh stay in sync with
manifest.json. When a new cloud provider is added, CLAUDE.md mandates
updating both files with endpoint mappings, auth env vars, and API
dispatchers. These tests catch configuration drift automatically:
- ALL_RECORDABLE_CLOUDS completeness and no duplicates
- get_endpoints(), get_auth_env_var(), call_api() coverage parity
- _strip_api_base() URL patterns match fixture directories
- Fixture directories have required _env.sh and _metadata.json
- Auth env vars in record.sh match manifest auth fields
- Shell script conventions (shebang, set -eo pipefail, no echo -e)
- Test infrastructure conventions (NO_COLOR, cleanup traps, counters)
Agent: test-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace fragile blocklist validation and printf '%q' escaping in upload_file()
with strict allowlist regex [a-zA-Z0-9/_.~-]+ across all non-SSH cloud providers.
For codesandbox, additionally migrate from shell command interpolation to SDK
filesystem API via environment variables, eliminating the injection surface entirely.
Affected clouds: codesandbox, daytona, e2b, fly, koyeb, modal, northflank,
railway, render, sprite
Fixes#989
Agent: security-auditor
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract platform-specific install logic from monolithic installer functions
into small, focused helpers. Both functions had nested OS/package-manager
cascades (depth 3-4) that made the control flow hard to follow.
ensure_jq (shared/common.sh):
- Extract _install_jq_brew, _install_jq_apt, _install_jq_dnf, _install_jq_apk
- Extract _report_jq_not_found for the fallthrough error message
- Main function becomes a clean dispatcher + verification
ensure_gh_cli + _install_gh_binary (shared/github-auth.sh):
- Extract _install_gh_brew, _install_gh_apt, _install_gh_dnf
- Extract _detect_gh_platform, _fetch_gh_latest_version, _download_and_install_gh
- _install_gh_binary drops from 71 to 12 lines as a clean orchestrator
- ensure_gh_cli drops from 57 to 29 lines
No behavior changes. All tests pass, bash -n passes.
Agent: complexity-hunter
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Previously, Phase 3 fix agents only got the last 10 lines grepped from
the log file per failing script. This was often insufficient to diagnose
the root cause. Now runs `bash test/mock.sh {cloud}` per failing cloud
and feeds the complete output to the fix agent.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Both mock.sh and record.sh now run each cloud's tests/recordings
concurrently as background jobs instead of sequentially.
Results are aggregated after all clouds finish.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
When a cloud's fixture recording fails 3+ consecutive QA cycles, the
system now auto-creates a GitHub issue flagging the persistent failure.
This catches stale API keys, changed endpoints, and other silent
regressions that would otherwise go unnoticed.
- Persistent tracker at .docs/qa-record-failures.json (git-ignored)
- Counter increments on failure, resets on success
- Deduplicates: skips issue creation if one already exists for that cloud
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
The bot was under-updating test/mock.sh when adding new clouds because
the prompt only mentioned URL stripping. Now lists all 4 required
mock.sh functions and all 5 required record.sh functions explicitly.
Also adds a "Mock Test Infrastructure" reference table to CLAUDE.md so
both human contributors and bots know exactly what to update.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
- Validate SPAWN_HOME is an absolute path, reject relative paths to prevent
unintended file writes (addresses #980)
- Resolve SPAWN_HOME to canonical form to collapse .. segments
- Strip __proto__, constructor, and prototype keys from parsed manifest JSON
to prevent prototype pollution (addresses #979)
- Apply sanitization to all manifest ingestion paths (GitHub fetch, disk cache,
local dev manifest)
- Add 12 tests covering path validation and JSON sanitization
Agent: security-auditor
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add interactive confirmation prompt before clearing spawn history
(spawn list --clear) to prevent accidental data loss
- Show total prompt length in dry-run preview when prompt exceeds 100
characters, so users can verify the correct prompt was loaded
- Add "Rerun previous" suggestion to non-interactive terminal fallback
- Show "(shown first)" hint when clouds with credentials are detected
in interactive picker, so users understand the sort order
- Add repository URL to spawn version output for discoverability
Agent: ux-engineer
Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
* fix: strip ANSI colors before grepping test summary
The mock test output uses ANSI escape codes for colored ✓/✗/━━━
characters, so the grep in the Post summary step couldn't match
them. Strip colors with sed first.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: use NO_COLOR standard instead of sed to strip ANSI codes
mock.sh now respects the NO_COLOR env var (https://no-color.org/).
CI sets NO_COLOR=1 so grep matches ✓/✗/━━━ cleanly.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
The QA cycle was auto-merging stale QA PRs that were mergeable.
Now it only closes stale ones — merging is left for human review.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
