@codacy/verity-cli 0.35.0-experimental.ce90a7b → 0.35.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +74 -0
- package/bin/verity.js +13546 -11352
- package/data/skills/verity-init/SKILL.md +59 -9
- package/package.json +3 -10
package/CHANGELOG.md
CHANGED
|
@@ -64,6 +64,80 @@ installs are untouched.
|
|
|
64
64
|
which is the ordinary case for someone whose first Verity install *is* the
|
|
65
65
|
plugin.
|
|
66
66
|
|
|
67
|
+
## [0.35.0] — 2026-10-30
|
|
68
|
+
|
|
69
|
+
**A new first run: sign in, answer two questions, and work with your team's Standard.**
|
|
70
|
+
|
|
71
|
+
### 🚀 `verity init`
|
|
72
|
+
|
|
73
|
+
- **Sign-in comes first.** Then init asks two questions: when Verity reviews
|
|
74
|
+
(each turn, each commit, each push; all three by default), and, when the
|
|
75
|
+
service writes project rules for you, whether to send a profile of the
|
|
76
|
+
repository for them.
|
|
77
|
+
- **Your team's Standard.** When the service has one, init offers it, or a new
|
|
78
|
+
one (`--standard use|new`). It never pushes a Standard over a team's Standard
|
|
79
|
+
that it could not read or write.
|
|
80
|
+
- **"Not finished" means it.** When no usable Standard could be set up, init
|
|
81
|
+
says why and exits 1, so scripts and the plugin stop there.
|
|
82
|
+
- **Trivy is the static analysis**, configured by Verity
|
|
83
|
+
(`.verity/static-analysis.json`), on new and existing projects: secrets,
|
|
84
|
+
malicious packages, and critical and high vulnerabilities. ESLint and
|
|
85
|
+
Semgrep no longer run.
|
|
86
|
+
- **The analyzer comes with Verity.** Its first run on a machine downloads
|
|
87
|
+
Trivy's vulnerability database (about 1.5 GB on disk). Init starts it in
|
|
88
|
+
the background. A review waits for the analyzer at most 45 s, then says
|
|
89
|
+
that static analysis did not run.
|
|
90
|
+
- **Re-running init removes a `.codacy/` config that an older Verity
|
|
91
|
+
wrote**, and stages the removal. It prints the command that restores it.
|
|
92
|
+
- **Formal verification is on for new projects.** The Lean toolchain downloads
|
|
93
|
+
in the background.
|
|
94
|
+
- **Telemetry** is one question, for the whole machine.
|
|
95
|
+
|
|
96
|
+
### 📐 Standards
|
|
97
|
+
|
|
98
|
+
- **`verity standard pull`** writes your team's Standard into
|
|
99
|
+
`.verity/standard.yaml`, and keeps a copy of your file.
|
|
100
|
+
- **Your own fields stay.** A field that Verity does not read is kept as you
|
|
101
|
+
wrote it, at every level of the Standard.
|
|
102
|
+
- **`verity standard push` sends only a Standard.** The file is checked before
|
|
103
|
+
anything is sent.
|
|
104
|
+
- **`verity config moments`** sets when Verity reviews.
|
|
105
|
+
|
|
106
|
+
### 📊 Usage
|
|
107
|
+
|
|
108
|
+
- **`verity usage`** shows your review budgets and your team's reviews. **`verity
|
|
109
|
+
upgrade`** opens the plans page.
|
|
110
|
+
- **A limit never blocks your work.** When a plan limit stops a review, Verity
|
|
111
|
+
says which limit and when it resets, and the commit or the push goes on.
|
|
112
|
+
|
|
113
|
+
### 🔧 Fixes
|
|
114
|
+
|
|
115
|
+
- Includes security and privacy fixes.
|
|
116
|
+
|
|
117
|
+
## [0.34.4] — 2026-10-26
|
|
118
|
+
|
|
119
|
+
**A formal finding is checked before it blocks a push, and formal looks for security bugs.**
|
|
120
|
+
|
|
121
|
+
### 🛡️ Formal verification
|
|
122
|
+
|
|
123
|
+
- **A reproduced counterexample is checked before it blocks.** A replay proves
|
|
124
|
+
that a failing order runs against stubs of your code. It does not prove that
|
|
125
|
+
the order can happen in production. Verity now investigates each reproduced
|
|
126
|
+
counterexample with the read-only tools of the agent review. It follows the
|
|
127
|
+
code to its entry point and looks for what prevents the order: a lock, a
|
|
128
|
+
unique constraint, a transaction, an idempotency key. Two independent checks
|
|
129
|
+
run, and the more cautious answer governs. A finding is lowered only when
|
|
130
|
+
every quote of the check appears in your code: PREVENTED and ARTIFACT show as
|
|
131
|
+
checked and never block, and REAL-LOW is reported without blocking. A check
|
|
132
|
+
that does not finish leaves the finding as it was.
|
|
133
|
+
`VERITY_FORMAL_VERIFY=off` turns the check off.
|
|
134
|
+
- **Formal looks for security bugs.** The ranker gives each target a lens
|
|
135
|
+
(security or reliability) and up to three concrete hypotheses. A security
|
|
136
|
+
target is modelled from the attacker's side: isolation between tenants,
|
|
137
|
+
revocation, flow binding, event order, and check-then-use.
|
|
138
|
+
- **Generation sees the push.** The lines that your push changed are marked in
|
|
139
|
+
the code that formal models, so it looks first for a bug that the change made.
|
|
140
|
+
|
|
67
141
|
## [0.34.3] — 2026-10-21
|
|
68
142
|
|
|
69
143
|
**`verity usage` shows the reviews of each member of your org.**
|