feat: add /compare-models and /compare-models-free slash commands (t168)

.agents/scripts/commands/compare-models-free.md

@@ -0,0 +1,68 @@
+---
+description: Compare AI model capabilities using offline embedded data only (no web fetches)
+agent: Build+
+mode: subagent
+---
+
+Compare AI models using only embedded reference data. No web fetches, no API calls.
+Useful when working offline or to avoid token spend on web fetches.
+
+Target: $ARGUMENTS
+
+## Instructions
+
+1. Run the helper script to get structured model data:
+
+   ```bash
+   ~/.aidevops/agents/scripts/compare-models-helper.sh list
+   ```
+
+2. If specific models were requested, compare them:
+
+   ```bash
+   ~/.aidevops/agents/scripts/compare-models-helper.sh compare <model1> <model2> ...
+   ```
+
+3. If a `--task` was specified, get recommendations:
+
+   ```bash
+   ~/.aidevops/agents/scripts/compare-models-helper.sh recommend "<task>"
+   ```
+
+4. For pricing overview:
+
+   ```bash
+   ~/.aidevops/agents/scripts/compare-models-helper.sh pricing
+   ```
+
+5. For capability matrix:
+
+   ```bash
+   ~/.aidevops/agents/scripts/compare-models-helper.sh capabilities
+   ```
+
+6. **Do NOT fetch any web pages.** All data comes from the helper script's embedded database.
+   Note the "Last updated" date in the output so the user knows data freshness.
+
+7. Present results in a structured comparison table with:
+   - Pricing per 1M tokens (input and output)
+   - Context window sizes
+   - Capability matrix
+   - Task suitability recommendations
+   - aidevops tier mapping (haiku/flash/sonnet/pro/opus)
+
+## Examples
+
+```bash
+# Compare specific models (offline)
+/compare-models-free claude-sonnet-4 gpt-4o
+
+# Get recommendation for a task (offline)
+/compare-models-free --task "summarization"
+
+# Show all pricing (offline)
+/compare-models-free --pricing
+
+# Show capabilities matrix (offline)
+/compare-models-free --capabilities
+```

.agents/scripts/commands/compare-models.md

@@ -0,0 +1,59 @@
+---
+description: Compare AI model capabilities, pricing, and context windows (with optional live data)
+agent: Build+
+mode: subagent
+---
+
+Compare AI models by capability, pricing, context window, and task suitability.
+
+Target: $ARGUMENTS
+
+## Instructions
+
+1. Read `tools/ai-assistants/compare-models.md` for the full comparison workflow.
+
+2. Run the helper script to get structured model data:
+
+   ```bash
+   ~/.aidevops/agents/scripts/compare-models-helper.sh list
+   ```
+
+3. If specific models were requested, compare them:
+
+   ```bash
+   ~/.aidevops/agents/scripts/compare-models-helper.sh compare <model1> <model2> ...
+   ```
+
+4. If a `--task` was specified, get recommendations:
+
+   ```bash
+   ~/.aidevops/agents/scripts/compare-models-helper.sh recommend "<task>"
+   ```
+
+5. **Live data enrichment** (this is the full `/compare-models` command):
+   - Fetch latest pricing from provider documentation pages
+   - Cross-reference against embedded data
+   - Note any pricing changes since last update
+
+6. Present results in a structured comparison table with:
+   - Pricing per 1M tokens (input and output)
+   - Context window sizes
+   - Capability matrix
+   - Task suitability recommendations
+   - aidevops tier mapping (haiku/flash/sonnet/pro/opus)
+
+## Examples
+
+```bash
+# Compare specific models
+/compare-models claude-sonnet-4 gpt-4o gemini-2.5-pro
+
+# Get recommendation for a task
+/compare-models --task "code review"
+
+# Show all pricing
+/compare-models --pricing
+
+# Compare by tier
+/compare-models --tier medium
+```