The agent→model registry in route.sh is now the single source of truth for
Ollama Cloud model IDs. route.sh publishes the full registry as a step
output (agents=...), and install-opencode.sh derives the ollama-cloud
provider models: map from it (unprefixed IDs), instead of hardcoding a
second copy. The two files can no longer drift silently — adding/changing
a cloud model in the registry automatically updates the provider config.
- route.sh: emit agents=$(jq -c '.' /tmp/agents.json) to GITHUB_OUTPUT
- agent.yml: pass steps.prep.outputs.agents as AGENTS env to install-opencode.sh
- install-opencode.sh: derive CMODELS from $AGENTS; // {} keeps it valid
when no agent uses an ollama-cloud model (non-ollama repos)
Rebased onto the per-agent skill-scoping change so PR #25 carries both:
- route.sh keeps the registry 'skills' allow-list and emits skills as a step output
- install-opencode.sh writes the permission.skill block (deny-all + allow listed)
Pure refactor otherwise: each step's shell moves to its own file, called via
bash "$SCRIPTS/<name>.sh". The two extracted SKILL.md bodies are byte-identical to
main; routing/config/publish behavior is unchanged. Because this is a reusable
workflow (workflow_call) the runtime checkout is the caller's repo, so agent.yml now
checks THIS repo out into .agents-workflow/ (pinned @main) and points $SCRIPTS there.
senior
lead