Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add assafkip/kipi-system/plugin install prd-osWrote 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/commands/assafkip/kipi-system/prd-personas)<a href="https://agentmods.dev/commands/assafkip/kipi-system/prd-personas"><img src="https://agentmods.dev/badge/commands/assafkip/kipi-system/prd-personas/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/commands/assafkip/kipi-system/prd-personas"><img src="https://agentmods.dev/badge/commands/assafkip/kipi-system/prd-personas.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 | $0.00012 | $0.01316 |
| Opus 5 | $0.00006 | $0.00658 |
| Sonnet 5 | $0.00002 | $0.00263 |
| Haiku 4.5 | $0.00001 | $0.00132 |
Grade A, and why
prd-personas scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Run the planning-personas Skeptic session against the active PRD. Refuses if the PRD is not in draft state, and writes the result as a ## Persona Review section in the PRD spec.
Phase 0 runtime gate (ENFORCED)
The command refuses to run unless q-system/memory/working/prd-personas-baseline.md contains the literal line phase_0_kill_criterion: failed. Until Phase 0 has been measured and failed the 50% reduction target, this command is dead code. Use the commented-out Persona Review template in plugins/prd-os/templates/prd.md instead.
The check is the first thing the command does:
grep -q '^phase_0_kill_criterion: failed$' q-system/memory/working/prd-personas-baseline.md || { echo "refused: Phase 0 gate not cleared. Use the template scaffold instead." >&2; exit 2; }
This is enforcement, not a prose warning.
Steps
- Resolve the active PRD:
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/prd_runner.py" status
Capture the prd_id, spec_path, and status from the JSON. If no PRD is active, stop and tell the author to run /prd-start first. If the status is not draft, refuse with the current state and a hint to run /prd-revise if the PRD needs to come back to draft.
-
Read the PRD body. Read
${CLAUDE_PLUGIN_ROOT}/personas/skeptic.mdto load the pinned questions and anti-patterns. Do not paraphrase the questions; they are asked verbatim. -
Run the Skeptic session in conversation with the founder:
- Ask each of the three pinned questions in order.
- Wait for an answer to each before moving on.
- A brief answer is acceptable. A dismissive answer ("nothing important") fires the anti-pattern warning from
skeptic.md; surface the warning and re-ask the question once. After one re-ask, accept whatever the founder writes. - Skip is allowed: if the founder types only
skip(case-insensitive, optionally with trailing punctuation), the Q-A pair records the answer literally asskippedrather than as the founder's prose. This preserves the question in the durable section while honoring the founder's choice not to engage with that specific question. - The founder may type
abortorcancelat any time. In that case do NOTHING to the PRD spec; partial answers are discarded. - Interruption (session ends before all three questions answered) is treated the same as abort: nothing is written.
- Answer sanitization (ENFORCED): before writing answers to the PRD spec, reject any answer that contains a level-2 markdown heading (
^##at the start of any line) OR a fenced code block (```). These would corruptprd_split.py's extraction logic, which anchors on## Issuesand the first fence after it. If a founder answer contains either, refuse and re-ask the question with a hint to remove level-2 headings and fenced code blocks. The Q-A pair is only persisted after the sanitization check passes.
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.
- 10d ago First seen · 67 lines · 12 tokens per session scan A 1af02f13fd62
prd-personas is a command published in the GitHub repository assafkip/kipi-system (110 stars, last pushed yesterday), licensed MIT. It adds 12 tokens to every session and 1,316 once invoked, about $0.0001 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.