You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Fix installer runtime compatibility, tool bindings, and readiness - #289
The reported macOS/Linux installs could accept an incompatible Node release, finish runtime installation despite errors, or leave selected network tools absent from the native MCP registry. Artifact checks could then look successful even when discovery failed.
This builds on merged PRs #287 (host preflight/Zabbix) and #288 (dotenv). It adds shared Node/npm bootstrap policy, user-owned npm prefix handling, thirteen selected native registrations for the twelve reported components, and shared native/skill launch bindings to recorded interpreters and source paths. The installer now checks structured discovery and configured Ollama model availability/capabilities and records explicit configuration-required/unverified stages. It never treats an empty catalog with exit 0 as ready.
Existing credentials, custom registrations, persona files and approval controls are preserved. pyATS keeps the owned private loopback HTTP bridge. Source patch version: 1.6.2.
Validation
Full isolated unit suite: 826 passed, 2 skipped (94 test files).
Installed OpenClaw 2026.7.1-2 / Node 24.19.0 discovered Packet Buddy (12), Subnet Calculator (1), TTS (2), and NetBox (4) through generated registrations in disposable HOME/config/Python environments.
Actual local calculator canary returned the expected TEST-NET /30 result. Negative fixtures cover empty/failed/timed-out discovery, missing runtime/registration/credentials, and Ollama endpoint/model failures.
Boundaries
No operator gateway or device was changed. The original live CML/pyATS endpoints, six unnamed macOS failures, and arbitrary model answer grounding remain unverified. Hermes's saved registrations/direct stdio discovery are checked; native agent discovery remains explicitly unverified. Model capability advertisement does not prove a completed tool-calling agent turn.
Spec and evidence: specs/147-installer-followups/. Recovery: docs/INSTALLER-READINESS.md. Constitution review: existing safety/change-control and credential boundaries preserved; inventory totals unchanged by moving previously external declarations into native registration; docs/spec/tests updated. Milestone blog draft included, unpublished.
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/gtrace-path-analysis/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/netbox-reconcile/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/nmap-network-scan/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/nvd-cve/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/packet-analysis/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/pyats-network/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/rfc-lookup/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/subnet-calculator/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
workspace/skills/wikipedia-research/SKILL.md
⚠️Error: ⚠ tessl skill review is deprecated; forwarding to tessl review run.
✘ Please authenticate with Tessl to continue. Run `tessl login` to sign up or log in.
To improve your score, point your agent at the Tessl optimization guide. Need help? Jump on our Discord.
Feedback
Report issues with this review at tesslio/skill-review, or send private feedback from your terminal with tessl feedback.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem and behavior
The reported macOS/Linux installs could accept an incompatible Node release, finish runtime installation despite errors, or leave selected network tools absent from the native MCP registry. Artifact checks could then look successful even when discovery failed.
This builds on merged PRs #287 (host preflight/Zabbix) and #288 (dotenv). It adds shared Node/npm bootstrap policy, user-owned npm prefix handling, thirteen selected native registrations for the twelve reported components, and shared native/skill launch bindings to recorded interpreters and source paths. The installer now checks structured discovery and configured Ollama model availability/capabilities and records explicit configuration-required/unverified stages. It never treats an empty catalog with exit 0 as ready.
Existing credentials, custom registrations, persona files and approval controls are preserved. pyATS keeps the owned private loopback HTTP bridge. Source patch version: 1.6.2.
Validation
Boundaries
No operator gateway or device was changed. The original live CML/pyATS endpoints, six unnamed macOS failures, and arbitrary model answer grounding remain unverified. Hermes's saved registrations/direct stdio discovery are checked; native agent discovery remains explicitly unverified. Model capability advertisement does not prove a completed tool-calling agent turn.
Spec and evidence:
specs/147-installer-followups/. Recovery:docs/INSTALLER-READINESS.md. Constitution review: existing safety/change-control and credential boundaries preserved; inventory totals unchanged by moving previously external declarations into native registration; docs/spec/tests updated. Milestone blog draft included, unpublished.