Automated README summary
Automatically summarized from the README on 28 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
pi-warden is a TypeScript guardrail layer for the Pi coding agent. It checks writes against project rules, holds hard-to-undo actions, catches unverified "done" claims and stuck loops, and trims tool output, reporting to the agent rather than the user.
How it uses Jev
Jev scores each command, write, and edit for irreversibility (0.9+ holds, 0.5–0.9 warns) and judges writes against rules in pi-warden.md. After a hold, a follow-up request reads the user's reply and the agent's turn, releasing only when both approved and reply_points_at_action score 0.7+. Verdicts return as held tool results or short steers.
Primitives:scorechoice
Technique worth stealing
Returning the model's verdict to the agent as a held tool result or steer, so the agent self-corrects without a user dialog.
Try it
pi install npm:pi-warden, then /warden enable with a TypeSafe key and /warden init.
Model judgments from README excerpts
Project signals use repository metadata, an existing summary and the first 700 README characters, not the full project. “Not found” means not detected in that material; it does not establish absence. These are the original judgments with recorded models and dates, not source audits or run results.
Models & evaluation dates
- Inclusion gate
jev-1.14.0- Project signals
jev-1.14.0- Category
jev-1.14.0- Pattern tags
jev-1.14.0
Evidence
These signals use repository metadata, an existing summary and at most 700 cleaned README characters. ≥ 0.60 means reported in that excerpt; ≤ 0.40 means not found there. The middle band is inconclusive. None is a source-code audit or a test result. The runnable-tool question excludes research experiments; the recommendation question asks for a learning shortlist, not a quality verdict.
- Jev is core to the projectFound in material0.61
- Shows a System One patternunclear0.51
- Describes handling uncertaintynot found in the excerpt0.04
- Reports measured numbersFound in material0.98
- Is a runnable toolFound in material0.99
- Top-10 learning shortlist signalnot found in the excerpt0.34
- Is a replica of the modelnot found in the excerpt0.05
- Solves a common problem (0–2)score on a 0–2 scale1.28 / 2
- About JevFound in material0.96
no-judgment band (0.40–0.60)
no-judgment band (0.40–0.60)
The Jev model and date for each judgment stage are listed above. Card written by DeepSeek V4.1 Flash from the README on 28 Oct 2026.