Under author review · MIT · 31 tests · 6,415 assertions
Retrieval Context Provenance Auditor
Payload-free provenance for every retrieval an agent uses, so a bad decision can be traced back to its sources.
Press room · facts, sources, downloads
SHPBL gives coding agents executable software capability plus the governed evidence to use it honestly. Twelve SHPBL-built agent-security tools are published free under MIT; one is currently submitted to Tenable’s CyberAgents Exchange and the rest are under the author’s own review before resubmission.
12
Public MIT security tools
1
Currently submitted to the Exchange
0
Listings published today
MIT
Public tool licences
The news in one paragraph
The agent-evidence series addresses retrieval provenance, tamper-evident action records, cross-agent delegation, behavioural drift, synthetic canary evidence, and proof-gated defense evolution. Each is an independently usable, zero-runtime-dependency MCP server with public tests and an MIT-licensed implementation. SHPBL composed the inventions from its own governed capability library and CMPSBL, then published clean-room implementations from written specifications.
Submission status
Counterfactual Immune Forge is submitted to the Exchange as pull request #190. No SHPBL listing has ever been merged or is published there today. The other eleven tools are held under the author’s own review before submission. Nothing here implies Tenable review, approval, certification or endorsement.
Under author review · MIT · 31 tests · 6,415 assertions
Payload-free provenance for every retrieval an agent uses, so a bad decision can be traced back to its sources.
Under author review · MIT · 60 tests · 158 assertions
An append-only, hash-linked evidence ledger for agent actions, verifiable from the exported file alone.
Under author review · MIT · 71 tests · 127 assertions
Signed delegation between agents where scope can only narrow and a handoff can only be used once.
Under author review · MIT · 28 tests · 118 assertions
Detects when an agent's behaviour quietly changes against a signed baseline of normal.
Under author review · MIT · 25 tests · 72 assertions
Synthetic-only canary markers plus a payload-free evidence chain proving where they reappeared.
Submitted · MIT · 44 tests
Proof-gated defense evolution: a fix is promoted only against the same sealed attack that beat it.
Independently checkable record
Exchange · current state
1 submitted · 0 merged
Counterfactual Immune Forge is open as pull request #190. No SHPBL listing has been merged or published on the Exchange. The other eleven tools remain under the author’s own review before submission.
Verify on GitHub ↗GitHub · repository commits
6 commits
The account’s commit history remains on the public contributor graph for tenable/cyberagents-exchange.
Copy-ready
SHPBL is governed software-capability infrastructure for coding agents. It audits a repository as it exists, identifies reusable capability, constrains work through explicit authority and evidence gates, and preserves approved results for later runs. The agent remains the engineer; SHPBL supplies executable capability, provenance, licensing, verification, and retention discipline.
Kenneth E. Sweet Jr. is the founder and sole developer of SHPBL. His work focuses on governed software reuse, evidence-preserving agent workflows, deterministic runtime claims, and software that can be inherited. He publishes twelve free MIT agent-security tools with their repositories, tests, specifications, provenance records, and negative findings open for inspection.
Approved media

SHPBL crystal mark
Square brand mark · PNG · 512 × 512
Download original ↓
SHPBL social card
Landscape brand card · PNG · 1200 × 630
Download original ↓
Kenneth E. Sweet Jr.
Founder portrait · JPEG
Download original ↓
SHPBL product artwork
Agent shipping plate · JPEG
Download original ↓Interviews, verification, technical questions
For interviews, source verification, product demonstrations, or higher-resolution assets, email dev@KESJr.com.