fix(jtk)!: shape-stable fulltext, fail-closed sprint resolution, deterministic automation export

skills/Jira/CliReference.md

@@ -29,7 +29,7 @@ Prompts for: Atlassian instance URL, email, API token (from https://id.atlassian
 
 | Flag | Description |
 |------|-------------|
-| `--fulltext` | Disable truncation of descriptions and comments (`--no-truncate` is a deprecated alias kept during migration) |
+| `--fulltext` | Disable truncation of descriptions and comments without changing list output shape |
 | `--id` | Emit only the primary identifier (useful for scripting). Takes precedence over `--fulltext` |
 | `--no-color` | Disable colored output |
 | `-v, --verbose` | Enable verbose output |
@@ -150,7 +150,7 @@ Common transition names: "To Do", "In Progress", "In Review", "Done" (instance-d
 
 ## Sprints
 
-`--board` and sprint positional arguments accept either a numeric ID or a name (resolved via cache — see SKILL.md Cache Warming).
+`--board` and sprint positional arguments accept either a numeric ID or a name (resolved via cache — see SKILL.md Cache Warming). Sprint names must resolve uniquely; ambiguity reports candidate IDs, while unresolved names require a cache refresh or numeric ID.
 
 | Command | Description |
 |---------|-------------|

skills/Jira/SKILL.md

@@ -87,7 +87,7 @@ Use `jtk refresh --status` to inspect cache freshness without fetching anything.
 `jtk` commands follow a text-first output model (per the repo's [Output Artifact Contract](https://github.com/open-cli-collective/atlassian-cli/blob/main/docs/ARTIFACT_CONTRACT.md)):
 
 - **`--fields <csv>`** — explicitly selects supported fields on commands that expose field projection.
-- **`--fulltext`** — disables truncation of descriptions and comments. Use when the user needs full body content (e.g., "show the full description"). `--no-truncate` is a deprecated alias kept during the migration; prefer `--fulltext`.
+- **`--fulltext`** — disables truncation of descriptions and comments without changing list output shape. Use when the user needs full body content (e.g., "show the full description").
 - **`--id`** — emits only the primary identifier (issue key, account ID, etc.) and takes precedence over `--fulltext` and `--fields` (projection is bypassed entirely). Stdout is identifier-only, one per line, safe to pipe; any continuation notice goes to stderr.
 
 `automation export` is the only resource command that emits JSON (`set-credential --json` is the control-plane exception). For every other command, use `--id` for scripting composition.

skills/Jira/Workflows/ManageComments.md

@@ -20,7 +20,7 @@ View and add comments on Jira issues.
 jtk comments list PROJ-123
 ```
 
-Optional: `--max N` to control page size (default 50). Use `--fulltext` global flag to disable truncation of long comment bodies.
+Optional: `--max N` to control page size (default 50). Use `--fulltext` to disable truncation of long comment bodies without changing the table shape.
 
 ### Add Comment
 
@@ -55,7 +55,7 @@ The `COMMENT_ID` is a numeric ID (e.g., `12345`). If the user refers to the comm
 
 ## Output Format
 
-- **List comments:** Show each comment with author, timestamp, and body
+- **List comments:** A table with ID, author, timestamp, and body; `--fulltext` preserves these columns
 - **Add comment:** Confirm comment was added with issue key and a snippet of the comment text
 
 ## Post-Action

tools/jtk/CHANGELOG.md

@@ -11,6 +11,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - **Breaking:** Removed global `--extended` and `issues list/search --all-fields`. Use explicit `--fields`; `--fulltext` controls truncation only. ([#455](https://github.com/open-cli-collective/atlassian-cli/issues/455))
 - **Breaking:** Pagination and continuation prose now goes to stderr, keeping `--id` stdout identifier-only. `--max` is a positive single-page size; issue list/search no longer auto-exhaust pages. ([#455](https://github.com/open-cli-collective/atlassian-cli/issues/455))
+- **Breaking:** `comments list --fulltext` now preserves table shape and only disables truncation; deprecated `--no-truncate` aliases were removed. Sprint-name filters now fail closed on ambiguous or unresolved names. Automation export now validates and deterministically formats API JSON. ([#455](https://github.com/open-cli-collective/atlassian-cli/issues/455))
 - **Breaking:** Short alias for `--file` renamed from `-f` to `-F` on `attachments add`, `automation create`, and `automation update`. `-f` continues to mean `--field` on field-setting commands (`issues create`/`update`, `transitions do`). No back-compat alias. ([#339](https://github.com/open-cli-collective/atlassian-cli/issues/339))
 - Default page size for paginated commands converged to 50: `issues list` and `issues search` were 25, `users search` was 10. `users search` and `dashboards list` also gain the `-m` short alias for `--max`. ([#340](https://github.com/open-cli-collective/atlassian-cli/issues/340))
 - `links types`, `issues types`, `boards list`, `sprints list`, and `users get` now serve from the local instance cache by default — removes per-command API calls in the most common paths. Run `jtk refresh` to update. ([#328](https://github.com/open-cli-collective/atlassian-cli/pull/328), [#329](https://github.com/open-cli-collective/atlassian-cli/pull/329), [#330](https://github.com/open-cli-collective/atlassian-cli/pull/330))

tools/jtk/README.md

@@ -293,6 +293,8 @@ jtk issues list --project MYPROJECT --fields summary,status,customfield_10005
 | `--fields` | | | Comma-separated display columns (headers, Jira field IDs, or human names) |
 | `--next-page-token` | | | Token for next page of results |
 
+Sprint names must resolve uniquely from the cache; ambiguous names report candidate IDs, and unresolved names require a cache refresh or numeric sprint ID.
+
 ---
 
 ### `jtk issues get <issue-key> [issue-key...]`
@@ -757,6 +759,8 @@ jtk comments list PROJ-123 --fields ID,AUTHOR
 | `--fulltext` | | `false` | Show full comment bodies without truncation (global) |
 | `--fields` | | | Comma-separated display fields |
 
+`--fulltext` preserves the table columns and row shape; it only disables body truncation.
+
 **Arguments:**
 - `<issue-key>` - The issue key (**required**)
 
@@ -1215,12 +1219,13 @@ jtk automation export 123 > rule-backup.json
 
 | Flag | Default | Description |
 |------|---------|-------------|
-| `--compact` | `false` | Output minified JSON |
+| `--compact` | `false` | Output whitespace-normalized minified JSON |
 
 **Arguments:**
 - `<rule-id>` - The rule ID (**required**)
 
 > Note: Output is always JSON — this is the only resource command that emits JSON directly (the control-plane `set-credential --json` envelope is the other exception).
+> Invalid JSON returned by the API fails with empty stdout.
 
 ---
 

tools/jtk/internal/cmd/OUTPUT_SPEC.md

@@ -98,6 +98,8 @@ All entity-reference flags (`--assignee`, `--project`, `--board`, `--sprint`, li
 - No match + looks like a raw ID → pass through unchanged
 - No match + looks like a name → fail with suggestion to `jtk refresh <resource>`
 
+Sprint names must resolve to exactly one cached sprint ID. Ambiguous names list candidate IDs; unresolved names fail with refresh and numeric-ID guidance. `issues list --sprint` never falls back to broader name-based JQL.
+
 ```
 $ jtk issues assign MON-4820 "John Smith"
 Ambiguous user "John Smith" — 3 matches:
@@ -295,18 +297,11 @@ ID | AUTHOR | CREATED | BODY
 21242 | Aaron Wong | 2026-04-16 | Short audit conclusion after the current code changes: The major source-level accessibility findings on CapOne-specific surfaces appear to be addressed or materially improv...
 ```
 
-**`comments list MON-4810 --fulltext`:** one block per comment:
+**`comments list MON-4810 --fulltext`:** the same table with body truncation disabled:
 ```
-ID: 21242
-Author: Aaron Wong
-Created: 2026-04-16
-Body:
-Short audit conclusion after the current code changes:
-The major source-level accessibility findings on CapOne-specific surfaces
-appear to be addressed or materially improved:
-- loading / redirect states now expose accessible status messaging
-- the unsupported-package modal now exposes both title and description correctly
-...
+ID | AUTHOR | CREATED | BODY
+21242 | Aaron Wong | 2026-04-16 | Short audit conclusion after the current code changes:
+The major source-level accessibility findings on CapOne-specific surfaces appear to be addressed or materially improved: ...
 ```
 
 ### `links`
@@ -406,7 +401,11 @@ Description: Creates Tasks when a new Onboarding Epic is created
 
 **`automation get <id> --show-components`:** dumps the full component tree as indented text (trigger → conditions → actions).
 
+<<<<<<< HEAD
+**`automation export <id>`:** validates and emits the rule definition as pretty-printed JSON to stdout. This is the round-trip format consumed by `automation create --from-file`. `--compact` emits whitespace-normalized minified JSON. Invalid API JSON is an error and emits nothing. This command bypasses the global flag system.
+=======
 **`automation export <id>`:** emits the rule definition as pretty-printed JSON to stdout. This is the round-trip format consumed by `automation create --file`. `--compact` minifies. This command bypasses the global flag system.
+>>>>>>> 455-jtk-pagination
 
 ### `dashboards`
 

tools/jtk/internal/cmd/automation/export.go

@@ -19,8 +19,8 @@ func newExportCmd(opts *root.Options) *cobra.Command {
 		Short: "Export automation rule as JSON",
 		Long: `Export the full automation rule definition as JSON.
 
-This outputs the exact JSON returned by the API, suitable for editing
-and re-importing via 'jtk auto update'. Output is always JSON.
+This validates and formats the JSON returned by the API, suitable for
+editing and re-importing via 'jtk auto update'. Output is always JSON.
 
 RECOMMENDED WORKFLOW:
   jtk auto export <rule-id> > rule.json
@@ -51,17 +51,14 @@ func runExport(ctx context.Context, opts *root.Options, ruleID string, compact b
 		return err
 	}
 
+	var buf bytes.Buffer
 	if compact {
-		_, err = fmt.Fprintln(opts.Stdout, string(raw))
-		return err
+		err = json.Compact(&buf, raw)
+	} else {
+		err = json.Indent(&buf, raw, "", "  ")
 	}
-
-	// Pretty-print the JSON
-	var buf bytes.Buffer
-	if err := json.Indent(&buf, raw, "", "  "); err != nil {
-		// If indenting fails, output raw
-		_, err = fmt.Fprintln(opts.Stdout, string(raw))
-		return err
+	if err != nil {
+		return fmt.Errorf("formatting automation rule JSON: %w", err)
 	}
 
 	_, err = fmt.Fprintln(opts.Stdout, buf.String())

tools/jtk/internal/cmd/automation/export_test.go

@@ -0,0 +1,78 @@
+package automation
+
+import (
+	"bytes"
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/open-cli-collective/atlassian-go/testutil"
+
+	"github.com/open-cli-collective/jira-ticket-cli/api"
+	"github.com/open-cli-collective/jira-ticket-cli/internal/cmd/root"
+)
+
+func TestRunExportFormatsJSONDeterministically(t *testing.T) {
+	t.Parallel()
+	tests := []struct {
+		name    string
+		raw     string
+		compact bool
+		want    string
+	}{
+		{"pretty input defaults to pretty", "{\n    \"name\": \"rule\",\n    \"enabled\": true\n}", false, "{\n  \"name\": \"rule\",\n  \"enabled\": true\n}\n"},
+		{"compact input defaults to pretty", `{"name":"rule","enabled":true}`, false, "{\n  \"name\": \"rule\",\n  \"enabled\": true\n}\n"},
+		{"compact normalizes whitespace", " { \"name\" : \"rule\", \"enabled\" : true } ", true, `{"name":"rule","enabled":true}` + "\n"},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			server := automationExportServer(tt.raw)
+			defer server.Close()
+
+			opts, stdout := exportTestOptions(t, server)
+			testutil.RequireNoError(t, runExport(context.Background(), opts, "rule-1", tt.compact))
+			if stdout.String() != tt.want {
+				t.Fatalf("output mismatch:\n got %q\nwant %q", stdout.String(), tt.want)
+			}
+		})
+	}
+}
+
+func TestRunExportMalformedJSONLeavesStdoutEmpty(t *testing.T) {
+	t.Parallel()
+	for _, compact := range []bool{false, true} {
+		server := automationExportServer(`{"broken":`)
+		opts, stdout := exportTestOptions(t, server)
+		err := runExport(context.Background(), opts, "rule-1", compact)
+		server.Close()
+
+		if err == nil {
+			t.Fatalf("compact=%v: expected malformed JSON error", compact)
+		}
+		if stdout.Len() != 0 {
+			t.Fatalf("compact=%v: stdout must stay empty, got %q", compact, stdout.String())
+		}
+	}
+}
+
+func automationExportServer(raw string) *httptest.Server {
+	return httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if r.URL.Path == "/_edge/tenant_info" {
+			_, _ = w.Write([]byte(`{"cloudId":"test-cloud"}`))
+			return
+		}
+		_, _ = w.Write([]byte(raw))
+	}))
+}
+
+func exportTestOptions(t *testing.T, server *httptest.Server) (*root.Options, *bytes.Buffer) {
+	t.Helper()
+	client, err := api.New(api.ClientConfig{URL: server.URL, Email: "t@x.com", APIToken: "tok"})
+	testutil.RequireNoError(t, err)
+	stdout := &bytes.Buffer{}
+	opts := &root.Options{Stdout: stdout, Stderr: &bytes.Buffer{}}
+	opts.SetAPIClient(client)
+	return opts, stdout
+}

tools/jtk/internal/cmd/comments/comments.go

@@ -7,8 +7,6 @@ import (
 
 	"github.com/spf13/cobra"
 
-	"github.com/open-cli-collective/atlassian-go/present"
-
 	"github.com/open-cli-collective/jira-ticket-cli/api"
 	"github.com/open-cli-collective/jira-ticket-cli/internal/cmd/root"
 	jtkpresent "github.com/open-cli-collective/jira-ticket-cli/internal/present"
@@ -42,7 +40,6 @@ func Register(parent *cobra.Command, opts *root.Options) {
 func newListCmd(opts *root.Options) *cobra.Command {
 	var maxResults int
 	var nextPageToken string
-	var noTruncate bool
 	var fieldsFlag string
 
 	cmd := &cobra.Command{
@@ -61,14 +58,12 @@ func newListCmd(opts *root.Options) *cobra.Command {
 			return jtkpresent.ValidateMax(maxResults)
 		},
 		RunE: func(cmd *cobra.Command, args []string) error {
-			return runList(cmd.Context(), opts, args[0], maxResults, nextPageToken, noTruncate || opts.IsFullText(), fieldsFlag)
+			return runList(cmd.Context(), opts, args[0], maxResults, nextPageToken, opts.IsFullText(), fieldsFlag)
 		},
 	}
 
 	cmd.Flags().IntVarP(&maxResults, "max", "m", 50, "Page size")
 	cmd.Flags().StringVar(&nextPageToken, "next-page-token", "", "Decimal startAt for the next page")
-	cmd.Flags().BoolVar(&noTruncate, "no-truncate", false, "Show full comment bodies without truncation")
-	_ = cmd.Flags().MarkDeprecated("no-truncate", "use --fulltext instead")
 	cmd.Flags().StringVar(&fieldsFlag, "fields", "", "Comma-separated display fields (labels)")
 
 	return cmd
@@ -92,9 +87,6 @@ func runList(ctx context.Context, opts *root.Options, issueKey string, maxResult
 	var selected []projection.ColumnSpec
 	var projected bool
 	spec := jtkpresent.CommentListSpec
-	if noTruncate {
-		spec = jtkpresent.CommentDetailSpec
-	}
 	if !idOnly {
 		selected, projected, err = projection.Resolve(
 			ctx,
@@ -133,33 +125,19 @@ func runList(ctx context.Context, opts *root.Options, issueKey string, maxResult
 		return jtkpresent.Emit(opts, model)
 	}
 
-	optional := projection.HasOptionalFields(selected, spec)
-	var model *present.OutputModel
-	if noTruncate {
-		model = jtkpresent.CommentPresenter{}.PresentListFullWithPagination(result.Comments, optional, hasMore, nextToken)
-		if projected {
-			projectAllDetailSectionsInModel(model, selected)
-		}
-	} else {
-		model = jtkpresent.CommentPresenter{}.PresentListWithPagination(result.Comments, optional, hasMore, nextToken)
-		if projected {
-			projection.ApplyToTableInModel(model, selected)
-		}
+	model := jtkpresent.CommentPresenter{}.PresentListWithPagination(
+		result.Comments,
+		projection.HasOptionalFields(selected, spec),
+		noTruncate,
+		hasMore,
+		nextToken,
+	)
+	if projected {
+		projection.ApplyToTableInModel(model, selected)
 	}
 	return jtkpresent.Emit(opts, model)
 }
 
-// projectAllDetailSectionsInModel rewrites every DetailSection of model
-// to the selected fields, leaving non-Detail sections (e.g. the
-// pagination MessageSection) untouched.
-func projectAllDetailSectionsInModel(model *present.OutputModel, selected []projection.ColumnSpec) {
-	for i, s := range model.Sections {
-		if ds, ok := s.(*present.DetailSection); ok {
-			model.Sections[i] = projection.ProjectDetail(ds, selected)
-		}
-	}
-}
-
 // commentsHasMore computes pagination using the authoritative API metadata,
 // falling back to a full-page heuristic when Total is unavailable (Jira Cloud
 // occasionally returns Total=0).