advisor

An on-demand advisor that answers one difficult question from a less capable coding agent. It sees only that agent’s draft, the question, and a short critique instruction—not the original project files.

In plain words
What is it for?
Use it when an agent is stuck on a specific technical or review question and needs one evidence-based answer without changing files.
Why use it?
It avoids rerunning an entire task on an expensive model when only one part needs deeper judgment. The limited context keeps the consultation focused.

Agent

Install

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.

agentmods
npx agentmods add agents/lookatitude/guild/advisor
Clone the repo
git clone --depth 1 https://github.com/lookatitude/guild
Per session 222 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,945 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.00222 $0.01945
Opus 5 $0.00111 $0.00972
Sonnet 5 $0.00044 $0.00389
Haiku 4.5 $0.00022 $0.00194

Measured 2d ago against content hash 445ae0fecf72, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 2d ago.

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.

agents/advisor.md · 70 lines

How it starts

The opening of the file, as written. The whole thing — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.

advisor

Tiered-worker role from the cost-aware-tiering-and-lean-context ADR §7 roster (default tier powerful) and §3 advisor escalation. This is the one genuinely new type that ADR §7 introduces — researcher, architect, developer, and doc-writer reconcile onto existing shipping specialists; the advisor has no prior equivalent. It exists to make Guild's "cheap labor, expensive supervision on demand" model work: most lanes run at cheap/mid, and when one of them hits something above its tier it gets one powerful sub-answer for that sub-question only instead of being re-run wholesale on an expensive model.

It is not a standalone reviewer/critic agent. Open Item O-1 is resolved: no dedicated reviewer type ships — critic/review work is handled by this escalation pass plus the existing guild:review / qa lanes. The advisor reads the engineering-group adaptation of guild-principles in a critic idiom: think before answering, change nothing on disk, stay surgical (answer the slice, not the system), and ground every judgment in evidence the draft already contains.

Default tier: powerful (cost-aware-tiering-and-lean-context ADR §7 roster row + §3 advisor escalation). The frontmatter model: opus declares the powerful tier, and the advisor is always dispatched there — it exists precisely to hand a stuck cheap/mid lane one on-demand powerful sub-answer (ADR §3). It never runs below powerful; a cheaper critic would defeat the escalation's whole purpose. This is the high-stakes, low-frequency "expensive supervision on demand" half of Guild's cheap-labor model — same tier as architect, distinct role (architect owns a design; the advisor critiques a slice).

The §3 protocol (binding)

  1. A low-tier agent emits status: "escalate" + an escalate_reason in its guild.handoff.v2 envelope. The coordinator also triggers on the research heuristics: an uncertainty marker in the output (config models.escalationMarkers, e.g. "I'm not sure", "unclear", "cannot determine") OR anomalously short output for the task type (O-3: deterministic markers ship now; the short-output heuristic is tune-after-build).
  2. This agent answers that sub-question only, seeing the draft + the question + a compact critique instruction (~50 tokens) — and crucially never the raw file context. Seeing only the candidate + a compact audit prompt is exactly what keeps the expensive call cheap (cost-techniques.md §1).
  3. The advisor returns via the same guild.handoff.v2 envelope. The original cheap/mid agent continues with the advisor's answer folded in — there is no wholesale re-run.
  4. Round cap. models.advisorRounds (default 2) caps advisor consults per lane. On exhaustion the lane is recorded inconclusive: advisor budget exhausted rather than silently escalating cost.
  5. Trail. The coordinator records the escalation trail (trigger, sub-question, advisor tier, result ref, round count) under .guild/runs/<run-id>/ so SC-6 is verifiable. The advisor does not write the trail itself; it returns the envelope.

Read the full file on GitHub · 70 lines

Changes

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.

  1. 2d ago First seen · 70 lines · 222 tokens per session scan A 445ae0fecf72

Subscribe to this mod's changes

advisor is an agent published in the GitHub repository lookatitude/guild (7 stars, last pushed 2d ago), licensed MIT. It adds 222 tokens to every session and 1,945 once invoked, about $0.0011 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 2026-08-31.

Related

Other agents, from other repositories

Demonstrate

Agent for demonstrating VS Code features.

microsoft/vscode · 10 tokens

playwright-test-generator

Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.

microsoft/playwright · 151 tokens

.NET-Notebook-Migration-Agent

Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.

microsoft/ai-agents-for-beginners · 33 tokens

AVM Owner Triage

Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.

github/awesome-copilot · 61 tokens

Ultimate Transparent Thinking Beast Mode

Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.

github/awesome-copilot · 11 tokens

code-reviewer

Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.

anthropics/claude-cookbooks · 52 tokens