diff --git a/PILOT.md b/PILOT.md
index 7d87761..5f80b9d 100644
--- a/PILOT.md
+++ b/PILOT.md
@@ -1844,3 +1844,8 @@ This is prospective instrumentation hardening, not a retrospective explanation o
 The triage API and CLI accept optional fixed-enum relative diagnostic costs for supplied candidate IDs. Only locally verified low/medium/high/unknown tokens may reach Jev. Costs guide next-check ordering and are explicitly not causal likelihood evidence. Default requests, local resolution, original failing exits, non-execution and mandatory validation remain unchanged. CLI rejects unknown values, duplicate costs and costs for unsupplied candidates before backend use.
 
 Six API tests, including changed-cost cache invalidation, three new CLI integration tests, ten existing triage CLI tests and the full 848-test suite pass (40.532 s). Graph refreshed (3465 nodes, 7016 edges, 235 communities). Hosted CI remains required before merge. README documents source-only scope; the published version is unchanged. No real-trial speed or quality benefit is claimed.
+
+
+## VCR407 — Coding-skill guidance for diagnostic costs
+
+The bundled English coding workflow explains optional caller-verified relative check costs, omission of unverified estimates, direct execution of decisive local checks and separation from causal ranking. It preserves required test execution and local fallback. This documents the merged API/CLI without claiming measured benefit or changing the published release. Nine skill-pack tests and three coding-workflow tests pass. Hosted CI remains required before merge.
diff --git a/TASKS.md b/TASKS.md
index 8f7403c..66574de 100644
--- a/TASKS.md
+++ b/TASKS.md
@@ -1248,3 +1248,8 @@ This is prospective instrumentation hardening, not a retrospective explanation o
 The triage API and CLI accept optional fixed-enum relative diagnostic costs for supplied candidate IDs. Only locally verified low/medium/high/unknown tokens may reach Jev. Costs guide next-check ordering and are explicitly not causal likelihood evidence. Default requests, local resolution, original failing exits, non-execution and mandatory validation remain unchanged. CLI rejects unknown values, duplicate costs and costs for unsupplied candidates before backend use.
 
 Six API tests, including changed-cost cache invalidation, three new CLI integration tests, ten existing triage CLI tests and the full 848-test suite pass (40.532 s). Graph refreshed (3465 nodes, 7016 edges, 235 communities). Hosted CI remains required before merge. README documents source-only scope; the published version is unchanged. No real-trial speed or quality benefit is claimed.
+
+
+## VCR407 — Coding-skill guidance for diagnostic costs
+
+The bundled English coding workflow explains optional caller-verified relative check costs, omission of unverified estimates, direct execution of decisive local checks and separation from causal ranking. It preserves required test execution and local fallback. This documents the merged API/CLI without claiming measured benefit or changing the published release. Nine skill-pack tests and three coding-workflow tests pass. Hosted CI remains required before merge.
diff --git a/src/jevcompass/bundled_skills/jevcompass-coding-workflow/SKILL.md b/src/jevcompass/bundled_skills/jevcompass-coding-workflow/SKILL.md
index bf3e7c4..d2ffe7d 100644
--- a/src/jevcompass/bundled_skills/jevcompass-coding-workflow/SKILL.md
+++ b/src/jevcompass/bundled_skills/jevcompass-coding-workflow/SKILL.md
@@ -79,3 +79,7 @@ jevcompass triage --exit-code 1 --kind import \
 ```
 
 The CLI accepts enum values, not the failure text, code, path, or log. Add an observation flag only when that exact enum fact was checked locally. Use the returned locally authored step to guide your next investigation. A suggested step is not a confirmed cause or an automatic fix, and the original test failure remains a failure until the test is rerun and passes. If local inspection leaves only one plausible explanation, investigate it directly and skip triage.
+
+When relative check effort has been established locally and competing diagnostics remain useful, you may add repeatable `--diagnostic-cost HYPOTHESIS=COST` options for supplied hypothesis IDs. Costs are `low`, `medium`, `high`, or `unknown`. Omit unverified estimates; do not fabricate costs to steer the result. For example, append `--diagnostic-cost import_module_missing=low` only if the missing-package check is verified to be low effort in this task. Cost guides diagnostic order, not causal likelihood. A cheap decisive local check should be performed directly rather than requesting remote confirmation.
+
+If you opt into `--rank-hypotheses`, assess the complete causal order separately from the selected next diagnostic step. An inexpensive check is not evidence that its associated cause is more likely. Unknown costs, incomplete rankings or backend failure do not remove any required tests; continue with the normal local investigation.