docs: record AgentShield corpus benchmark evidence

Record AgentShield PR #60 corpus benchmark evidence in the ECC 2.0 GA roadmap and update the next AgentShield slice.

Validation:
- markdownlint roadmap
- npm test: 2324 passed
- harness audit: 70/70
- harness adapters: PASS, 11 adapters
- observability readiness: 14/14
- GitHub Actions matrix green
This commit is contained in:
Affaan Mustafa 2026-05-12 07:15:10 -04:00 committed by GitHub
parent a8c03ad350
commit 7a4c25f1df
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -55,6 +55,9 @@ As of 2026-05-12:
- AgentShield PR #59 added self-contained HTML executive summaries with risk - AgentShield PR #59 added self-contained HTML executive summaries with risk
posture, critical/high priority findings, category exposure, README/API posture, critical/high priority findings, category exposure, README/API
docs, built-CLI smoke validation, and 1,704-test coverage. docs, built-CLI smoke validation, and 1,704-test coverage.
- AgentShield PR #60 added category-level built-in corpus benchmark output,
a `readyForRegressionGate` signal, terminal `--corpus` category coverage,
README/API docs, built-CLI smoke validation, and 1,705-test coverage.
- ECC PR #1778 recovered the useful stale #1413 network/homelab architect-agent - ECC PR #1778 recovered the useful stale #1413 network/homelab architect-agent
concepts. concepts.
- ECC-Tools PR #26 added cost/token-risk predictive follow-ups for AI routing, - ECC-Tools PR #26 added cost/token-risk predictive follow-ups for AI routing,
@ -180,7 +183,7 @@ Acceptance:
- Supply-chain intelligence covers MCP package provenance and has an extension - Supply-chain intelligence covers MCP package provenance and has an extension
path for npm/pip reputation, CVEs, typosquats, and dependency risk. path for npm/pip reputation, CVEs, typosquats, and dependency risk.
- Prompt-injection corpus and regression benchmark are ready for continuous - Prompt-injection corpus and regression benchmark are ready for continuous
rule hardening. rule hardening with category-level coverage and regression-gate output.
- Enterprise reports include JSON plus self-contained HTML executive output - Enterprise reports include JSON plus self-contained HTML executive output
with risk posture, priority findings, and category exposure. with risk posture, priority findings, and category exposure.
@ -226,7 +229,7 @@ Acceptance:
## Next Engineering Slices ## Next Engineering Slices
1. Finish AgentShield prompt-injection corpus/regression benchmark work and 1. Decide whether AgentShield PDF export adds value beyond the merged HTML
decide whether PDF export adds value beyond the merged HTML executive report. executive report and corpus benchmark output.
2. Extend ECC Tools deep analysis and Linear/project sync without flooding the 2. Extend ECC Tools deep analysis and Linear/project sync without flooding the
workspace. workspace.