@pickled-dev/cli 0.17.0 → 0.17.2
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/README.md +2 -2
- package/dist/index.js +151 -144
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -164,10 +164,10 @@ URL sources are NOT scanned by the audit's trap cross-reference in v1; they are
|
|
|
164
164
|
|
|
165
165
|
## Toolsets
|
|
166
166
|
|
|
167
|
-
Matrix mode (`scenario.matrix.toolsets`) iterates each scenario across named toolset profiles.
|
|
167
|
+
Matrix mode (`scenario.matrix.toolsets`) iterates each scenario across named toolset profiles. Two profiles ship today:
|
|
168
168
|
|
|
169
169
|
- **`none`** (the deterministic baseline). Pickled injects the cell's active source content into the agent's prompt. Citation contract applies if `requiredSources` is declared. Same scoring shape as non-matrix scenarios.
|
|
170
|
-
- **`web`** on Claude Code only. Maps to `allowedTools: ["WebSearch", "WebFetch"]` on the cell's Claude Code target. Source is NOT injected; the cell's prompt is rewritten to name the active source as the discovery target ("the canonical source for this question is at ..."). Citation contract is skipped; the cell scores on traps + `expected.includes`/`excludes
|
|
170
|
+
- **`web`** on Claude Code only. Maps to `allowedTools: ["WebSearch", "WebFetch"]` on the cell's Claude Code target. Source is NOT injected; the cell's prompt is rewritten to name the active source as the discovery target ("the canonical source for this question is at ..."). Citation contract is skipped; the cell scores on traps + `expected.includes`/`excludes` + tool-use provenance. Tool-use provenance is a hard veto: a cell that does not invoke at least one of the configured web tools is forced to `NO` with confidence `0`, because an answer pulled from model prior knowledge cannot testify to the tool path the cell is meant to test.
|
|
171
171
|
|
|
172
172
|
Declare profiles at the top level of `pickled.yml`:
|
|
173
173
|
|