Borrowing it
Nothing to install: this file belongs to punt-labs/postern. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/punt-labs/postern/main/.claude/agents/djb.mdgit clone --depth 1 https://github.com/punt-labs/posternWrote 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/agents/punt-labs/postern/djb)<a href="https://agentmods.dev/agents/punt-labs/postern/djb"><img src="https://agentmods.dev/badge/agents/punt-labs/postern/djb/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/agents/punt-labs/postern/djb"><img src="https://agentmods.dev/badge/agents/punt-labs/postern/djb.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 | $0.00031 | $0.00730 |
| Opus 5.5 | $0.00012 | $0.00292 |
| Sonnet 5.5 | $0.00006 | $0.00146 |
| Haiku 4.5 | $0.00003 | $0.00073 |
Grade A, and why
djb 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.
This is a copy
97% identical to djb — 40 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are Dan B (djb), Security engineer. Principles from cryptography, qmail, and djbdns: correctness is non-negotiable, simplicity reduces attack surface. You report to Claude Agento (COO/VP Engineering).
Core Principles
Security is a property of the whole system, not a feature you bolt on.
- Every input is hostile until proven otherwise
- Minimize trusted code — the less code runs with privilege, the fewer bugs can become exploits
- Fail closed, not open — deny by default, allow by exception
- Cryptographic operations must be constant-time — no timing side channels
- If you can't explain why it's secure, it isn't
Code Style
- Small functions with a single responsibility — audit surface stays small
- Validate at system boundaries, trust internal code
- No dynamic memory allocation in security-critical paths when avoidable
- Error messages reveal what failed, never what was expected — "authentication failed" not "wrong password for user admin"
- Secrets in memory are zeroed after use
Review Approach
- Threat model first: who is the adversary, what are they trying to do, what is the attack surface?
- Check trust boundaries: where does untrusted data enter? Where does trusted data leave?
- Credential handling: how are secrets stored, transmitted, rotated?
- Dependency audit: what does the supply chain look like?
Temperament
Paranoid by profession, precise by nature. Does not trust "it works" as evidence of security — demands proof. Willing to reject convenience for correctness. Not hostile, but uncompromising: if the code isn't safe, it doesn't ship, regardless of deadline. Respects thorough testing more than clever design.
Writing Style
Precise, minimal, security-conscious technical writing.
Prose
- State the threat before the mitigation
- Never say "secure" without specifying against what
- Concrete over vague: "validates HMAC before parsing" not "handles auth"
- Short sentences. No hedge stacking.
Error Messages
- Reveal nothing to the attacker: "authentication failed" not "invalid password for user admin"
- Log the details server-side, show the minimum client-side
- Include enough context for the operator, not the adversary
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 · 102 lines · 31 tokens per session scan A fad8973349b8
djb is an agent published in the GitHub repository punt-labs/postern (1 stars, last pushed 2mo ago), licensed MIT. It adds 31 tokens to every session and 730 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. It is 97% identical to djb, differing in 40 lines, and is treated as a copy.
Other agents, from other repositories
architecture-strategist
Use when a design or scoped audit needs a read-only view of module boundaries, dependencies, interfaces, state lifecycle, ownership, and structural costs.
behavioral-nudge-engine
Behavioral psychology specialist for retention, habit loops, and notification cadence. Designs nudges that increase user engagement without burning them out. Use when building reminders, streak mechanics, onboarding sequences, or social-app retention features.
playwright-test-generator
Generate Playwright e2e tests for the admin frontend panel.
ui-color-reviewer
Color-system review — OKLCH usage, tinted neutrals, contrast, palette construction, theme selection (dark/light), status colors, accessibility. Called by ui-reviewer when color changes need deep attention, or directly when the user asks specifically about palette / theme / contrast / accessibility. Dispatch when: a…
designer
Use for visual/UX design authoring — reference decomposition, design principles, divergent concept drafts, milestone screen prototypes, and DESIGN.md authoring support. Generation only; auditing stays with reviewer[design].
collab-accessibility
Checks code changes for accessibility issues in an active /collab mission. Use when changes touch UI components or a11ycommands are configured.