Skip to content

Add auditing-agent-memory-with-as-of-system-time skill - #28

Open
something1703 wants to merge 1 commit into
cockroachlabs:mainfrom
something1703:add-skill/security-and-governance/auditing-agent-memory-with-as-of-system-time
Open

something1703 wants to merge 1 commit into
cockroachlabs:mainfrom
something1703:add-skill/security-and-governance/auditing-agent-memory-with-as-of-system-time

Conversation

@something1703

@something1703 something1703 commented Aug 18, 2026 •

Copy link
Copy Markdown

Reconstructing exactly what an AI agent's memory contained at a past instant using AS OF SYSTEM TIME, for audit/deposition/compliance review: the historical-read pattern, the "silently reads current state instead of historical state" failure mode (and how to make it fail loudly instead), the garbage-collection window that bounds how far back a historical read can reach, extending gc.ttlseconds for retention obligations, and the access-control tension between historical reads and erasure requests.

Distilled from building a temporal-recall/deposition feature (explain(action_id) reconstructing exactly what an agent believed at decision time) on CockroachDB, generalized as a reusable skill.

What does this PR do?

Related Issue

Closes #

Type of Change

  • New skill
  • Skill update/improvement
  • Documentation update
  • Infrastructure/tooling change
  • Bug fix

Skill Compliance Checklist

  • SKILL.md includes required frontmatter fields (name, description)
  • Skill name follows naming conventions (lowercase, hyphens, gerund form preferred)
  • Skill name matches directory name exactly
  • Description is specific and includes "when to use" trigger keywords
  • Description is max 1024 characters
  • Skill is under 500 lines (or uses references/ for detailed content)
  • References official CockroachDB documentation (not duplicate content)
  • Includes safety guardrails for risky operations (if applicable)
  • Tested with at least one AI agent (manual validation)
  • No time-sensitive information (version numbers, dates, "currently")
  • Directory structure follows specification (only scripts/, references/, assets/ subdirs)
  • No reserved words in skill name ("anthropic", "claude")
  • Local validation passes: python scripts/validate-spec.py skills/

Documentation Updates

  • README.md
  • CONTRIBUTING.md
  • Skill SKILL.md files
  • Other (specify):

Testing

Manual testing:

  • Tested with AI agent (specify which one):
  • Verified skill is discoverable with appropriate prompts
  • Validated technical accuracy against CockroachDB docs
  • Tested any scripts or commands included in the skill

Automated testing:

  • Local validation script passes: python scripts/validate-spec.py skills/
  • CI validation will run automatically on this PR

Additional Context

Reviewer Notes


By submitting this PR, I confirm:

Reconstructing exactly what an AI agent's memory contained at a past
instant using AS OF SYSTEM TIME, for audit/deposition/compliance
review: the historical-read pattern, the "silently reads current
state instead of historical state" failure mode (and how to make it
fail loudly instead), the garbage-collection window that bounds how
far back a historical read can reach, extending gc.ttlseconds for
retention obligations, and the access-control tension between
historical reads and erasure requests.

Distilled from building a temporal-recall/deposition feature
(explain(action_id) reconstructing exactly what an agent believed at
decision time) on CockroachDB, generalized as a reusable skill.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant