agentsclimarketplace

Sentinel config

Skill AndurilCode/sentinel/skills/sentinel-config

View or update Sentinel configurationFrom its SKILL.md

Install
npx -y skills add AndurilCode/sentinel --skill sentinel-config

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.7 KB, as published. Nobody here has run it

Sentinel Config

View or update the Sentinel configuration in the current repository.

What to do

  1. Check that .claude/sentinel/config.yaml exists. If not, tell the user to run /sentinel-init first and stop.

  2. Read .claude/sentinel/config.yaml.

  3. If the user asked to view the config, show the current values and explain what each one does. Stop here.

  4. If the user asked to change something, use the Edit tool to update the specific value(s) in the config file. Preserve all comments. Do not rewrite the entire file.

Configuration reference

KeyDefaultDescription
backendollamaLLM backend: ollama, claude, or copilot
modelgemma3:4bDefault model for evaluation (backend-specific)
backends.ollama.urlhttp://localhost:11434Ollama endpoint
backends.ollama.model(top-level model)Default Ollama model
backends.claude.modelhaikuDefault Claude model
backends.copilot.modelgpt-5-miniDefault Copilot model
ollama_urlhttp://localhost:11434Ollama endpoint (legacy; prefer backends.ollama.url)
timeout_ms5000Per-rule evaluation timeout
confidence_threshold0.7Minimum confidence to count as violation
max_parallel4Concurrent LLM calls
ollama_concurrency1Max concurrent Ollama HTTP calls (GPU-bound)
thinkfalseEnable /think mode (slower, more accurate)
fail_opentrueSkip rule on error vs block
content_max_chars800File content truncation in prompts
log_filenullJSONL telemetry path
rules_dirrulesRules directory (relative to config dir)
tool_map(see reference.md)Tool name → trigger type mapping
mcp_prefixmcp__Prefix for detecting MCP tool names
mcp_separator__Separator for parsing MCP server/tool from tool name
context.enabledtrueEnable session context accumulator
context.backend(top-level backend)LLM backend for accumulator
context.modelgemma3:4bModel for accumulator
context.min_events3Minimum events before accumulator update
context.lock_timeout_s30Max wait for GPU lock (Ollama only)
context.summary_max_words150Token budget for rolling summary
scribe.enabledtrueEnable the Scribe convention learning pipeline
scribe.backend(top-level backend)LLM backend for scribe
scribe.model(top-level model)Default model for all scribe steps
scribe.extraction_model(scribe.model)Override for extraction (reflect + learn)
scribe.synthesis_model(scribe.model)Override for validation + synthesis
scribe.guidancenullPriority guidance text for extraction
scribe.thinkfalseEnable /think mode for validation+synthesis
scribe.extraction_timeout_ms15000Timeout for extraction LLM calls
scribe.extraction_num_predict1000Max output tokens for extraction
scribe.synthesis_timeout_ms15000Timeout for validation+synthesis LLM calls
scribe.synthesis_num_predict1000Max output tokens for validation+synthesis
scribe.temperature0.1LLM temperature for all scribe calls
scribe.transcript_budget_chars4000Max compacted transcript size before truncation
scribe.thresholds.extraction_confidence0.7Minimum confidence to store an observation
scribe.thresholds.draft_confidence0.8Minimum confidence for learn mode drafts
scribe.sources.documentationtrueEnable doc scanning via /sentinel-learn
scribe.doc_globs[CLAUDE.md, AGENTS.md, ...]File patterns for /sentinel-learn
scribe.notification.max_age_days7Max age for draft notifications

Backend and model recommendations

BackendModelNotes
ollama (8 GB RAM)gemma3:4bDense, ~3 GB at Q4, local
ollama (16 GB RAM)gemma3:12bDense, best local accuracy for block rules
claudehaikuFast cloud model, good for most rules
copilotgpt-5-miniFast cloud model via Copilot CLI

Backends can be mixed: use Ollama globally for speed, override specific block rules with backend: claude for higher accuracy.

Next steps

After viewing or updating config, suggest these to the user:

  • /sentinel-stats — see how the current configuration affects rule performance
  • /sentinel-rule — create a new rule if the rules directory is empty
  • /sentinel-learn — scan documentation for conventions to enforce automatically

Keep looking

Skills are one crate of 608,475. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.