@gobing-ai/spur 0.4.9 → 0.4.10

package/plugins/sp/skills/session-review/SKILL.md

@@ -149,39 +149,23 @@ step 4 excludes. Use `n/a` for any value not supported by the measurement.
 
 ### Resolved issues
 
-| Issue | Root cause | Resolution | Evidence | Confidence |
-| --- | --- | --- | --- | --- |
+| Issue | Root cause | Resolution | Evidence |
+| --- | --- | --- | --- |
 
 Evidence names the tool result, verification command, or concrete repository state visible in the
 session. Do not list ordinary implementation steps as issues.
 
-**Confidence** is `HIGH`, `MEDIUM`, or `LOW` for that row's claim. The level is **derived from the
-check that produced the claim, never assessed from how the row reads**, and the cell carries that
-verifier (e.g. `HIGH — gate exit 0, /tmp/g0.log re-run this session`). Add the row's **falsifier**
-when the level is not `HIGH`: what would have made it fail, or what moves it up. The same ladder
-grades this report's Open-issues table.
-
-| Level | Verifies when | Not this |
-| --- | --- | --- |
-| `HIGH` | An executed check in this session produced the result and its raw output is reachable (command + exit code, test counts, query row, gate log). Re-runnable now. | The code or change "looks correct"; a claim asserted only in prose. |
-| `MEDIUM` | The claim is inferred from observed facts, rests on prior-session or delegated evidence, or names state that moved after the check (stale by construction). Name which of the three. | Presenting an inference as an observation. |
-| `LOW` | No executable check was run. Name the missing evidence **and** the command that would ground it. | Dropping the row instead of marking it. |
-
-Two guards against level inflation:
-
-- Evidence that cannot be re-run, or that is only the agent's own summary, **caps the row at
-  `MEDIUM`** — a re-run is what promotes it.
-- An all-`HIGH` multi-finding review is a signal the ladder was not applied, not a clean session.
-  Where every row genuinely is `HIGH`, the cells show distinct check results rather than repeated
-  phrasing.
-
-Any API/library statement in a row cites the installed source as `path:line` or a URL, and a claim
-the session could not ground is written `LOW` with the missing evidence named.
+**Evidence is the check, not a grade** (ADR-138). A row resolved by an executed check this session
+names the command and its result (e.g. `gate exit 0, /tmp/g0.log`). A row resting on inference,
+prior-session or delegated evidence says so. A row with no check run names the missing evidence
+**and** the command that would ground it — never drop the row instead. No HIGH/MEDIUM/LOW label:
+a self-assigned level is as easy to write as the claim. Any API/library statement in a row cites
+the installed source as `path:line` or a URL.
 
 ### Open issues and risks
 
-| Issue or risk | State | Evidence or confirmation needed | Confidence |
-| --- | --- | --- | --- |
+| Issue or risk | State | Evidence or confirmation needed |
+| --- | --- | --- |
 
 ### Process and environment improvements
 

package/plugins/sp/skills/source-driven-development/SKILL.md

@@ -55,27 +55,28 @@ Do **not** use this skill for:
 
 - **Your own code's correctness** — that is `sp:code-verification` (requirements/AC) and `sp:code-review`.
 - **Stress-testing a design decision** — that is `sp:doubt-driven-development` (artifact vs contract).
-- **Well-known, version-stable stdlib** you can state with HIGH confidence — verification has a cost;
+- **Well-known, version-stable stdlib** you have already checked — verification has a cost;
   spend it where being wrong is likely and expensive.
 
 ## The process
 
-### Step 1 — Classify your confidence before you write
+### Step 1 — Know where each fact came from before you write
 
-| Level | Meaning | Action |
-|---|---|---|
-| **HIGH** | Verified against the pinned version's docs/source **this session** | Generate; cite the source inline |
-| **MEDIUM** | Recalled from a prior session; plausibly stale | Verify before relying on it |
-| **LOW** | Memory only, no source in hand | Stop — verify first; never present as fact |
+| Source in hand | Action |
+|---|---|
+| Checked against the pinned version's docs/source **this session** | Generate; cite the source inline |
+| Recalled from a prior session; plausibly stale | Verify before relying on it |
+| Memory only, no source in hand | Stop — verify first; never present as fact |
 
-If you cannot honestly claim HIGH for a version-specific behavior, you are at MEDIUM or LOW — verify.
+Do not grade the claim with a confidence label (ADR-138): a label is as easy to write as the claim.
+The citation is the evidence.
 
 ### Step 2 — Go to the primary source, in priority order
 
 1. The **type signature / source** in the installed, pinned dependency (the ground truth for *this* build).
 2. The **official docs for the pinned version** (not the latest — the version in the lockfile).
 3. Authoritative references when docs are thin.
-4. Memory — **LOW only**, and never as the final authority.
+4. Memory — never the final authority; verify before relying on it.
 
 State the **version** inline whenever behavior is version-specific.
 
@@ -91,40 +92,36 @@ Generate the code and **cite the source** (doc URL + date, or `pkg@version` + th
 point of use or in the task's notes. A claim about external behavior without a citation is
 unverified — flag it as such rather than presenting it as fact.
 
-### Step 5 — Close with a verification verdict
-
-Before finishing, verify the citations for the API/library claims you emitted and close with one
-line: `Verification: HIGH|MEDIUM|LOW — <first failing check and deviation, or 'all checks passed'>`.
+### Step 5 — Check the citations before finishing
 
 - **Citation coverage** — every API/library claim in the output carries a source citation (doc URL + date, or `pkg@version` + file/symbol); no uncited claim is asserted as fact.
 - **Source authority** — each citation points at the pinned version's primary source (installed signature/source, or official docs for the lockfile version), not the latest docs or memory.
-- **Honest confidence** — the stated HIGH/MEDIUM/LOW matches the evidence: HIGH requires a check against the pinned source this session; MEDIUM/LOW items are labeled, and LOW is never presented as fact.
-- **HIGH** — all three checks pass. **MEDIUM** — coverage intact but some citations are off-version (latest docs for a pinned dependency) or secondary-source only. **LOW** — an API/library claim lacks any citation, or a LOW-confidence claim was asserted as fact.
 
-Deterministic backstop: the `cited-version-strings`, `cited-api-claims`,
-`cited-api-claims-confidence` and `cited-high-confidence` rules
+Report the first failing check and its deviation, if any. Do not append a confidence verdict.
+
+Deterministic backstop: the `cited-version-strings` and `cited-api-claims` rules
 (`.spur/rules/docs/sourced-api-claims.yaml`) flag markdown lines that assert a version string or a
-time-anchored API claim with no citation anchor on the same line, claim lines that carry no
-HIGH/MEDIUM/LOW level, and claim lines that claim HIGH with no anchor — the one level that asserts
-"verified this session", so its source must be on the line; MEDIUM (recalled) and LOW (memory only)
-need no anchor by contract and are not flagged. They are an info-severity advisory backstop for
-this contract and are deliberately NOT wired into `recommended-pre-check` — run them ad hoc with
-`spur rule run --file .spur/rules/docs/sourced-api-claims.yaml`. The checks above and the SECUA
-checklist remain the enforcement points.
+time- or version-anchored API claim (`deprecated since`, `removed in 0.9`, `fixers (since 0.3.4)`)
+with no citation anchor on the same line. They are an info-severity advisory
+backstop for this contract and are deliberately NOT wired into `recommended-pre-check` — run them ad
+hoc with `spur rule run --file .spur/rules/docs/sourced-api-claims.yaml`. Dated working records
+(`docs/report*`, `docs/dogfood`, `docs/plans`, `docs/analysis`) are excluded, as are the task
+archives: a version measured by a past run is not an authorable claim. The checks above and the
+SECUA checklist remain the enforcement points.
 
 ## Common Rationalizations
 
 | Rationalization | Reality |
 |---|---|
-| "I've used this API a hundred times — I know the signature." | APIs change across versions and your memory is a silent cache. Verify against the *pinned* version; recall is MEDIUM at best. |
+| "I've used this API a hundred times — I know the signature." | APIs change across versions and your memory is a silent cache. Verify against the *pinned* version; recall is not a source. |
 | "It compiled, so I used it right." | Compiling proves existence and types, not contract. Ordering, preconditions, and error modes are not checked by the compiler. |
 | "The latest docs say so." | You ship against the *pinned* version, not latest. Read the docs for the version in the lockfile. |
 | "Close enough — I'll fix it if it breaks." | Guessing at an external contract moves the cost to runtime/production. Verify before generating; it is cheaper here. |
-| "Verifying every call is too slow." | Verify what you are *not sure of*, at the confidence level you honestly hold. HIGH needs no re-check; LOW always does. |
+| "Verifying every call is too slow." | Verify what you have no source for. A fact checked this session needs no re-check; memory always does. |
 
 ## Red Flags
 
-- Generating version-specific code you would rate MEDIUM or LOW confidence, without opening a source.
+- Generating version-specific code from memory, without opening a source.
 - Presenting an API claim as fact with no citation (doc + date, or `pkg@version` + symbol).
 - Treating "it exists" as proof it is used correctly (the two questions collapsed into one).
 - Reading the *latest* docs for a dependency pinned to an older version.
@@ -136,8 +133,7 @@ checklist remain the enforcement points.
 - [ ] Every version-specific external behavior relied on was checked against the pinned version's source/docs **this session**.
 - [ ] Both questions answered — the API exists **and** is called within its contract.
 - [ ] Claims about external behavior carry a citation (source + date, or `pkg@version` + symbol); unverifiable ones are flagged, not asserted.
-- [ ] Confidence is stated honestly (HIGH/MEDIUM/LOW); nothing at LOW was presented as fact.
-- [ ] Run closes with the Step 5 verdict line: `Verification: HIGH|MEDIUM|LOW — <first failing check and deviation, or 'all checks passed'>`.
+- [ ] Nothing recalled from memory alone was presented as fact.
 
 ## See also
 

package/plugins/sp/skills/spur-cli/references/agent.md

@@ -341,9 +341,12 @@ spur agent usage --dry-run  # print would-be changes; write nothing
 spur agent usage --json
 ```
 
-Runs `codexbar usage --format json --provider all` once, writes the snapshot to
+Runs `codexbar usage --format json` once — the operator-enabled provider set, deliberately not an
+all-provider sweep (`--provider all` covers every provider configured in CodexBar, most of them
+uncredentialed, and its wall time is the slowest of all of them) — writes the snapshot to
 `~/.config/spur/agent-usage.json` (`captured_at`, `source`, `providers[]`, `raw`), maps providers
-to executors via `agent.executors[].agent` (or the model's `<provider>/` prefix), and records
+to executors via `agent.executors[].agent` (with the `antigravity-cli` → `antigravity` alias) or the
+model's `<provider>/` prefix, and records
 `owner: quota` availability observations that the standard drain applies — the single availability
 write path. A provider is exhausted when any `primary|secondary|tertiary` window reports
 `usedPercent >= 100`; the window name and `resetsAt` go into the observation reason. Per-provider

package/plugins/sp/skills/spur-cli/references/features.md

@@ -30,7 +30,7 @@ what* or *how to write a scenario*, this skill.
 | `refresh` | Rebuild INDEX + each feature `## Tasks` table from task edges (**docs only**; no status change) | `--feature <id>` `--all` `--folder` `--json` |
 | `check [id]` | Validate one feature / the tree; the 4-layer gate; `--fix` repairs structural findings in place | `--strict` `--fix` `--as <status>` `--folder` `--json` |
 | `sync [id]` | Align feature **lifecycle status** with linked task states (real transitions + guards) | `--all` `--dry-run` `--force` `--folder` `--json` |
-| `verify [id]` | The `feature-verification` pass: run the check set once and record the bound PASS/FAIL receipt — workflow-internal (`feature-lifecycle` runs it) | `--run-id <id>` `--cmd <command>` `--confidence <level>` `--ack-low-confidence` |
+| `verify [id]` | The `feature-verification` pass: run the check set once and record the bound PASS/FAIL receipt — workflow-internal (`feature-lifecycle` runs it) | `--run-id <id>` `--cmd <command>` |
 | `handoff [id]` | The `idea-pipeline` finalization: task ordering, roster refresh, handoff report — workflow-internal | `--run-id <id>` `--spur-bin <cmd>` |
 
 **`refresh` vs `sync` (do not conflate):**

package/plugins/sp/skills/spur-cli/references/workflows/authoring-workflows.md

@@ -138,7 +138,7 @@ Spur's host additionally registers (`registerSpurBuiltins`, `packages/app/src/wo
 
 | Kind | Capability |
 | ---- | ---------- |
-| Action | `agent.run` (model call via `spur agent run`; `executor: 'fleet'` dispatches into the agent fleet), `decide` (inline `id`/`method`/`question`/`choices`/`default`/`evidence` → outcome written to `resultFile`; catalog-id references are pending task 1094 — use `spur decision run` for catalog decisions), `command.gate`, `rule.check`, `doctor.probe`, `response.validate`, `proof.fingerprint`, `run.artifact`, `file.exists`, `file.read`, `file.read.into-var`, `http.request`, `hitl.confirm`, `hitl.select`, `hitl.input`, `shell` (streaming) |
+| Action | `agent.run` (model call via `spur agent run`; `executor: 'fleet'` dispatches into the agent fleet), `decide` (inline `id`/`method`/`question`/`choices`/`default`/`evidence` → outcome written to `resultFile`; catalog-id references are pending task 1094 — use `spur decision run` for catalog decisions), `command.gate`, `rule.check`, `doctor.probe`, `proof.fingerprint`, `run.artifact`, `file.exists`, `file.read`, `file.read.into-var`, `http.request`, `hitl.confirm`, `hitl.select`, `hitl.input`, `shell` (streaming) |
 | Guard | `shell` (env-aware), `contract-violation` |
 
 Order matters for both guards and conditions: **the first that passes wins.** Put the discriminating

package/plugins/sp/skills/spur-dev/references/dev-operations.md

@@ -590,12 +590,11 @@ host instructions does not add one. An opt-in gate is explicit: `gitmsg --check
      - `keepachangelog` (default): `## [<version>] - <date>` header, then category headings per the keepachangelog convention — `### Added` / `### Fixed` / `### Changed` / `### Removed` / `### Other` — mapped from conventional-commit types.
      - `simple`: flat bulleted list grouped by type heading (`### feat`, `### fix`, …).
   6. Emit: with `<file>`, insert the section into that file above its newest `## [` release heading (or after its title when none exists) and touch nothing else in the file; without `<file>`, print it to stdout.
-  7. **Verify claims before emitting — two checks, both required, on every claim bullet.** Classify first: an *API/library claim* is a bullet asserting what code, an API, a library, a dependency or a version does — skip merge/bookkeeping bullets that assert no such fact (a status move, a filing, a doc relocation). Then run both checks and append one HTML comment line directly under the bullet carrying both results: `<!-- src: <citation> · <HIGH|MEDIUM|LOW> -->`.
-     - **(a) Source citation — the anchor.** Internal claims (what a commit or the code does): confirm against the commit diff (`git show --stat <hash>`, reading hunks when the summary asserts behavior) or the artifact's presence in the tree; cite `path[:line]` — bare path for whole-file claims, `path:line` for a specific behavior or symbol. External claims (framework/library/version behavior): confirm against official docs for the installed version; cite the URL. An anchor you cannot produce fails the check — reword the bullet or drop the claim; a citation is not a citation until it names something a reader can open.
-     - **(b) Confidence level — exactly one in-rubric token.** HIGH = the diff/artifact confirms the summary as written; MEDIUM = the evidence is real but narrower than the summary; LOW = cannot be confirmed from available sources (append the reason in one clause). A missing level, an off-rubric token, or two levels on one bullet fails the check.
-     - **Defect precision decides what must be fixed.** Both checks are precise once a bullet is classified as a claim: an annotation missing its anchor or its level is unambiguously wrong and blocks emitting. Which bullets *are* claims is the heuristic half — resolve it by this step's rule, and for a genuinely borderline bullet say so in the bullet's body rather than annotating or skipping it silently.
-  8. Self-check before emitting: every short hash in `<since>..<until>` appears exactly once, and every claim bullet carries exactly one `<!-- src: … -->` comment holding exactly one anchor and exactly one of `HIGH`, `MEDIUM`, `LOW` — no unannotated claim bullets, no duplicate annotations, no doubled or off-rubric levels. Fix any miss, duplicate, or bad level; do not report a verdict.
-  9. Reply in at most three lines: the range, bullet counts per category plus verified-claim counts (`x HIGH / y MEDIUM / z LOW`), and the output target.
+  7. **Verify claims before emitting — one required check on every claim bullet.** Classify first: an *API/library claim* is a bullet asserting what code, an API, a library, a dependency or a version does — skip merge/bookkeeping bullets that assert no such fact (a status move, a filing, a doc relocation). Then run the check and append one HTML comment line directly under the bullet carrying its anchor: `<!-- src: <citation> -->`. No confidence level is recorded (ADR-138): the anchor is the evidence.
+     - **Source citation — the anchor.** Internal claims (what a commit or the code does): confirm against the commit diff (`git show --stat <hash>`, reading hunks when the summary asserts behavior) or the artifact's presence in the tree; cite `path[:line]` — bare path for whole-file claims, `path:line` for a specific behavior or symbol. External claims (framework/library/version behavior): confirm against official docs for the installed version; cite the URL. An anchor you cannot produce fails the check — reword the bullet or drop the claim; a citation is not a citation until it names something a reader can open.
+     - **Defect precision decides what must be fixed.** The check is precise once a bullet is classified as a claim: an annotation missing its anchor is unambiguously wrong and blocks emitting. A summary broader than its evidence is reworded to what the anchor shows. Which bullets *are* claims is the heuristic half — resolve it by this step's rule, and for a genuinely borderline bullet say so in the bullet's body rather than annotating or skipping it silently.
+  8. Self-check before emitting: every short hash in `<since>..<until>` appears exactly once, and every claim bullet carries exactly one `<!-- src: … -->` comment holding exactly one anchor — no unannotated claim bullets, no duplicate annotations. Fix any miss or duplicate; do not report a verdict.
+  9. Reply in at most three lines: the range, bullet counts per category plus the verified-claim count, and the output target.
 - **Invariants:** Bullets are derived from `git log` only; the claim-verification pass (step 7) adds evidence from commit diffs, the working tree, and official docs. No quality gates — no `bun run gate`/lint/typecheck/test/`spur-check`/rule runs, and no edits beyond the one inserted section (its verification comments included). The file is left uncommitted.
 
 ### 9. gitmsg

package/plugins/sp/skills/spur-dev/references/execution-worktree-landing.md

@@ -425,7 +425,7 @@ git checkout <base-ref> && git merge --no-ff -m "<prepared message>" <branch>
 #     git checkout <base-ref> && git merge --no-commit <branch>
 # 2. resolve source conflicts by hand; generated files are then regenerated with the project's
 #    generator, never hand-merged
-#    (this repo: bun run build:plugin-lib && bun run --filter @gobing-ai/spur build:bundle)
+#    (this repo: bun run --filter @gobing-ai/spur build:bundle)
 # 3. stage every resolved path and regenerated bundle through the guard — an unmerged or
 #    unstaged path makes step 5 abort, and a conflict-marked path exits 3 (resolve, then retry):
 #      spur workflow commit-guard stage --run "$RUN_ID" -- <resolved paths and regenerated bundles>