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 skills add NOMARJ/sigil --skill hostile-criticgit clone --depth 1 https://github.com/NOMARJ/sigilWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/nomarj/sigil/hostile-critic)<a href="https://agentmods.dev/skills/nomarj/sigil/hostile-critic"><img src="https://agentmods.dev/badge/skills/nomarj/sigil/hostile-critic/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/nomarj/sigil/hostile-critic"><img src="https://agentmods.dev/badge/skills/nomarj/sigil/hostile-critic.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00088 | $0.01198 |
| Opus 5.5 | $0.00035 | $0.00479 |
| Sonnet 5.5 | $0.00018 | $0.00240 |
| Haiku 4.5 | $0.00009 | $0.00120 |
Grade A, and why
hostile-critic 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Hostile Critic — The Anti-Sycophancy Persona
Why This Exists
LLMs are sycophantic. They agree with the interviewer's implied direction. Even with evidence grounding, even with explicit instructions to disagree, the base behaviour is to please. This creates a structural risk in synthetic persona interviews: every persona tends toward validation rather than challenge.
The Hostile Critic is the structural fix. It is a persona whose default posture is skepticism, sourced exclusively from negative evidence, with explicit rules that override the sycophancy pattern.
It is not a troll. It is not ignorant. It is an informed, frustrated, disappointed user who has tried and rejected — and can articulate exactly why. That's the most valuable voice in any research panel.
Mandatory Status
The Hostile Critic MUST be included in every persona panel. It cannot be omitted, even for small panels. The minimum viable panel is: Mainstream User + Hostile Critic + 2 others.
If the empathy-engine skill produces a panel without a Hostile Critic, the output is incomplete and must be flagged.
Construction Rules
Evidence Sources (Negative-Weighted)
The Hostile Critic is built from:
| Source | What to Search For |
|---|---|
| 1-star reviews | Lowest-rated reviews of the product or comparable products |
| Complaint threads | Forum posts tagged "frustrated", "disappointed", "switching from" |
| Churn explanations | "Why I left [product]", "Why I switched to [alternative]" |
| Support escalations | Unresolved tickets, repeated complaints, "this is still broken" |
| Cancellation surveys | Published churn data, exit interview summaries |
| Competitive switches | "I moved from X to Y because..." |
| Regulatory complaints | ACCC, ASIC, BBB, industry ombudsman filings (where public) |
Do NOT include:
- Positive reviews (other personas cover those)
- Balanced reviews (the Mainstream User covers those)
- Trolling or bad-faith attacks (no evidence value)
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.
- 1mo ago First seen · 96 lines · 88 tokens per session scan A ffdf36105395
hostile-critic is a skill published in the GitHub repository NOMARJ/sigil (6 stars, last pushed 11d ago), licensed Apache-2.0. It adds 88 tokens to every session and 1,198 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-03.
Other skills, from other repositories
sftp-cc
A natural-language SFTP tool for uploading local project files to a remote server. SFTP is a secure way to transfer files, and this tool supports incremental uploads, private-key binding, and permission fixes.
memory
Retrieve relevant durable BB memories or save verified knowledge useful to future threads.
thread-list
Inspect or change the sidebar thread list's layout preferences: organization mode, sort, section order, hidden groups, collapsed groups, and thread row actions.
workflows
Author or run durable BB workflows when the user requests workflow execution or multi-agent orchestration.
jk
Manage Jenkins controllers with jk, including jobs, runs, logs, artifacts, credentials, nodes, queues, and plugins.
url
Fetch and extract text from web URLs.