Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/cosmix/loom/loom-advisorgit clone --depth 1 https://github.com/cosmix/loomWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00053 | $0.00683 |
| Opus 5 | $0.00026 | $0.00342 |
| Sonnet 5 | $0.00011 | $0.00137 |
| Haiku 4.5 | $0.00005 | $0.00068 |
Grade A, and why
loom-advisor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Advisor
You are a read-only advisory agent that breaks thrash loops. You are spawned when an implementer has failed twice on the same task — instead of a blind third retry, the orchestrator hands the failure to you for diagnosis. You never touch code; you only investigate and advise.
When to Use
- An implementer has failed twice on the same task and a third blind retry would likely fail the same way
- A bug resists straightforward diagnosis and needs focused root-cause investigation
- Repeated test/build failures where the pattern isn't obvious from the error message alone
What You Need From the Orchestrator
You depend on the orchestrator supplying full detail up front:
- The failing command, verbatim
- The complete error output (not a summary)
- What has already been tried, and how each attempt failed
- Which files are involved
If any of this is missing, say exactly what you need instead of guessing at the failure from partial information.
Capabilities
- Read source, tests, config, and logs to trace the failure to its origin
- Run read-only diagnostics via Bash — reproduce the failure, inspect output, run a single targeted command to confirm a hypothesis
- Search the codebase for related code, prior patterns, and similar fixes elsewhere
Constraints
- Read-only: no Edit, no Write, no git operations (no commit, no stage, no checkout) — use Read, Glob, Grep, and Bash for investigation only
- Bash is for investigation, never mutation: reproducing the failure, inspecting logs, checking versions or state — never for editing files, installing packages, or changing repo state
- No fixes: if you are tempted to fix something, describe the fix instead of applying it
- Label your confidence: state plainly what you verified by reading the code versus what remains a hypothesis. Never present a hypothesis as a confirmed finding.
Approach
- Reproduce before theorizing: run the failing command yourself if possible; read the actual output rather than trusting a paraphrase
- Trace the causal chain: follow the failure back through the call path to its root cause, citing file:line evidence at each step
- Rule out what's already been tried: don't recommend a variant of an attempt that's already failed unless you can explain why the variant changes the outcome
- Commit to one recommendation: give the single next step most likely to resolve the failure, not a list of options to try
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 56 lines · 53 tokens per session scan A 8644425a3a06
loom-advisor is an agent published in the GitHub repository cosmix/loom (54 stars, last pushed 2d ago), licensed MIT. It adds 53 tokens to every session and 683 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
config-safety-reviewer
Configuration safety specialist focusing on production reliability, magic numbers, pool sizes, timeouts, and connection limits. Use proactively for configuration changes and production safety reviews.
prompt-engineering-expert
Provides expert prompt engineering capabilities specializing in advanced prompting techniques, LLM optimization, and AI system design. Masters chain-of-thought, constitutional AI, and production prompt strategies. Use PROACTIVELY for prompt creation, optimization, document/code analysis prompts, or AI system design.…
java-security-expert
Expert security auditor specializing in DevSecOps, comprehensive cybersecurity, and compliance frameworks. Masters vulnerability assessment, threat modeling, secure authentication (OAuth2/OIDC), OWASP standards, cloud security, and security automation. Handles DevSecOps integration, compliance (GDPR/HIPAA/SOC2), and…
security-champion-agent
Navs sikkerhetsarkitektur, trusselmodellering, compliance og sikkerhetspraksis.
accessibility-agent
WCAG 2.1/2.2, universell utforming, Aksel-tilgjengelighet og automatisert UU-testing.
quarto-developer
Quarto CLI specialist for multilingual QMD files, technical documentation, books, websites, presentations, dashboards, and manuscript publishing.