Install
$ agentstack add skill-mec07-claude-skills-modelrouting ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Model Routing
Route subagents to the right model based on task type. Opus for reasoning-heavy work, sonnet for code production, haiku for mechanical tasks. This is advisory with a strong default — deviate when context warrants it, but justify the deviation.
Model Assignment Table
| Task Type | Model | Rationale | |---|---|---| | Planning, architecture, brainstorming, design | opus | Best reasoning for complex decisions | | Code review | opus | Deep reasoning catches subtle issues | | Debugging / root cause analysis | opus | Complex reasoning needed | | Refactoring | opus | Cross-cutting changes need deep reasoning to avoid getting stuck | | Code implementation, feature work | sonnet | Fast and capable for writing code | | Test writing | sonnet | Follows patterns, doesn't need deepest reasoning | | Simple lookups, renames, file searches, formatting | haiku | Cheapest, fast enough for mechanical tasks |
When to Use
When dispatching any subagent via the Agent tool, consult the table above and set the model parameter accordingly.
Agent(model: "opus", prompt: "Review this PR for correctness...")
Agent(model: "sonnet", prompt: "Implement the user auth flow...")
Agent(model: "haiku", prompt: "Rename userId to accountId across these files...")
Deviation Policy
- Default to the table assignment
- Deviation is allowed when context warrants it (e.g., a trivial rename-style refactor might warrant sonnet instead of opus)
- When deviating, state why briefly in the agent prompt
- This skill applies only to subagent dispatch — the main conversation model is the user's choice
Self-Improvement
When a subagent produces a poor result and the likely cause is the model choice (e.g., haiku struggled with a task that needed deeper reasoning, or opus was overkill for something mechanical), update the Model Assignment Table in this file with the new scenario. This keeps the routing table comprehensive and grounded in real experience.
How to apply:
- Identify the task type that was misrouted
- Add a new row to the table, or refine an existing row's scope, to cover the scenario
- Include a brief rationale so future routing decisions are informed
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: mec07
- Source: mec07/claude_skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.