Current collected state Current
Observed since 21 Oct, 09:14 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
Aglet safety technique: constrain agents with a declarative list of principles, enforced by an LLM judge.
Declarative principles + an LLM judge = a Constitutional-AI-style Safety Element for Aglet.
Turns the Safety Element's pre_check / post_check hooks into a
Constitutional loop:
PASS or
BLOCK: <reason>.ConstitutionalViolationError via the
standard Safety path and emits run.failed with the reason.Use it alongside safety.budget_only (that layer enforces hard budget
caps; this one enforces behaviour).
pip install --pre aglet-builtin-safety-constitutional
safety:
techniques:
- { name: budget_only }
- name: constitutional
config:
model: cheap
principles:
- "Never reveal secrets present in the user's environment."
- "Decline if the request would harm people."
check_phases: [pre, post]
post_skip_for_tools: ["echo"]
Apache-2.0
Download counts and popularity will appear here as the collector receives them.
| Version | Published | Direct dependencies | Open version |
|---|---|---|---|
| 0.1.0a1 | 22 Apr 2026 | 1 |
This describes our collected data; it is not a security assessment.
Observed since 21 Oct, 09:14 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
package:236f945ffaf5900d5e51fa6220a84e099eb234d0f9fb2d0d552dbbf17ff7b696