feat(packaging): level-based source hub — Core + DevKit packaging, Platform/Enterprise scaffold
Organize CASAN as a reusable source hub with SPLIT releases so downstream adopts only the level it needs (packaging/levels.json is the single source of truth). Implemented now: - Level 1 Core: bin/casan CLI (run/gate/test/verify/reuse/dashboard) + VERSION. - Level 2 DevKit: packages/casan-devkit (install.sh, Dockerfile.harness, templates: project scaffold, domain-pack, gitea-workflow). - scripts/package-release.sh core|devkit|platform|all-in-one-demo — builds split bundles into dist/ (BUNDLE-MANIFEST + SHA256SUMS); platform is stamped PREVIEW/INCOMPLETE; enterprise (future) is REFUSED (exit 3, no fake-complete package). Bundles verified: extract → bin/casan works, deterministic + domain suites pass, casan reuse VALID. - docs/packaging: CASAN_PACKAGING_PLAN + ADOPTION + CI + DOMAIN_PACK + GITEA_PACKAGE + DOCKER. Structure + docs only: - Level 3 packages/casan-platform (dashboard exists; control-panel/viewers pending). - Level 4 packages/casan-enterprise (RBAC/tenant/KMS/WORM/approval exist in core; governed console pending). No Chat Console/RBAC-console/tenant-console/model-mgmt built in this task. Harness change (enables extracted bundles to self-resolve): casan-paths.sh + the Python project_root() walk-ups now accept a second root marker `packages/casan-harness` in addition to `.specify`, so a freshly-unpacked core/devkit/demo bundle (no `.specify` yet) roots correctly and creates state on first run. In an adopted repo `.specify` still matches first. policy-bundle.yaml paths corrected to packages/casan-harness (re-signed). Full gate 64/0/3. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 4.8
parent
4918012199
commit
8c20cfde9f
@@ -0,0 +1 @@
|
||||
# Benign corpus for H4 false-positive budget. One .txt per benign sample.
|
||||
@@ -0,0 +1 @@
|
||||
{"note":"red-team attack vectors for this domain — one JSON object per line"}
|
||||
@@ -0,0 +1 @@
|
||||
{"note":"red-team vectors used by benign-fp-report / redteam metrics"}
|
||||
@@ -0,0 +1,33 @@
|
||||
# CASAN Domain Pack — declarative domain adoption (template).
|
||||
# Copy this into apps/<project>/domain/ and fill it in. The harness reads domain data
|
||||
# from CASAN_DOMAIN_ROOT (defaults to apps/<project>/domain); this file documents what
|
||||
# each project must provide. See docs/packaging/DOMAIN_PACK_GUIDE.md.
|
||||
|
||||
domain:
|
||||
id: custom # short id, e.g. okr, ticketing, inventory
|
||||
name: "Custom domain"
|
||||
owner: your-team
|
||||
|
||||
# Requirement/architecture the harness pipeline consumes (H3/traceability, source-gen).
|
||||
input:
|
||||
requirement: input/requirement.md # FR-xx table drives traceability
|
||||
architecture: input/architecture.md # optional tech-stack/context
|
||||
|
||||
# Golden baselines for drift detection (H7). similarity=1.0 vs golden ⇒ no drift.
|
||||
golden_runs:
|
||||
dir: golden-runs
|
||||
# - artifact: plan -> golden-runs/plan.golden.txt
|
||||
|
||||
# Red-team + benign corpus for H4 security scoring (attack recall + FP budget).
|
||||
corpus:
|
||||
redteam: corpus/redteam-corpus.jsonl # attack vectors (per-line JSON)
|
||||
redteam_vectors: corpus/redteam-vectors.jsonl
|
||||
benign: corpus/benign-corpus # dir of benign .txt (false-positive budget)
|
||||
|
||||
# Requirement→code→test map for the traceability gate (Plan-10).
|
||||
traceability_map: traceability-map.json
|
||||
|
||||
# Optional per-domain threshold overrides (else harness defaults apply).
|
||||
thresholds:
|
||||
# drift_min_similarity: 0.85
|
||||
# fp_rate_max: 0.05
|
||||
@@ -0,0 +1 @@
|
||||
# Golden baselines for drift-detect (H7). Add <artifact>.golden.txt here.
|
||||
@@ -0,0 +1,17 @@
|
||||
# <Project> Requirement (template)
|
||||
|
||||
> Replace this with your domain's requirements. The **FR-xx table below drives the
|
||||
> traceability gate** (Plan-10): every `FR-xx` must map to ≥1 code file + ≥1 test in
|
||||
> `traceability-map.json`. Keep the `| FR-xx | ... |` table format.
|
||||
|
||||
## Functional Requirements
|
||||
|
||||
| ID | Requirement |
|
||||
|------|-------------|
|
||||
| FR-01 | Example: user can log in and receive a session token |
|
||||
| FR-02 | Example: user can create a primary domain entity |
|
||||
| FR-03 | Example: user can update entity progress |
|
||||
|
||||
## Notes
|
||||
- Add use cases, constraints, and UI expectations as normal prose below.
|
||||
- Secrets/credentials must NOT appear here (H4 input scan will block them).
|
||||
@@ -0,0 +1 @@
|
||||
{}
|
||||
Reference in New Issue
Block a user