What's wrong
~/.webcmd/skills/webcmd-adapter-author/SKILL.md:149 documents the recon step as:
Preferred: webcmd browser analyze <url> to get pattern, anti-bot signals, nearest adapter, and next step in one pass.
But the actual CLI requires a session argument first:
$ webcmd browser analyze --help
Usage: webcmd browser <session> analyze [options] <url>
Repro
Running the documented form fails immediately:
$ webcmd browser analyze "https://example.com/some/page"
✖ <session> is a required positional argument: webcmd browser <session> <command>
This is the very first command the skill's own runbook (Step 3) tells an agent to run, so any agent following the skill literally hits a failure before recon even starts.
Suggested fix
Either:
- Update the skill doc to
webcmd browser <session> analyze <url>, or
- Make
analyze tolerate a missing/default session the way browser init apparently does (that one worked fine without an explicit session in our testing).
Context
Found while a subagent built a new university-course-export adapter (Manchester) following webcmd-adapter-author end to end. Happy to share the full transcript excerpt if useful.
What's wrong
~/.webcmd/skills/webcmd-adapter-author/SKILL.md:149documents the recon step as:But the actual CLI requires a session argument first:
Repro
Running the documented form fails immediately:
This is the very first command the skill's own runbook (Step 3) tells an agent to run, so any agent following the skill literally hits a failure before recon even starts.
Suggested fix
Either:
webcmd browser <session> analyze <url>, oranalyzetolerate a missing/default session the waybrowser initapparently does (that one worked fine without an explicit session in our testing).Context
Found while a subagent built a new university-course-export adapter (Manchester) following
webcmd-adapter-authorend to end. Happy to share the full transcript excerpt if useful.