Two findings from Greptile review on PR #302:
1. tests/test_cli_v3.py:302 - The test asserted run_mock.call_args_list[0]
was the main runner's invocation, but fanout.run_competitor_fanout
submits main + competitors to a ThreadPoolExecutor and iterates with
as_completed. With zero-latency mocks, thread scheduling determines
which pipeline.run call lands first, so the competitor's call could
take index [0] and flake CI. Replace [0] indexing with a predicate
match on the canonicalized github_repos kwargs.
2. skills/last30days/scripts/last30days.py:662 - When auto_resolve returns
github_repos, it has already run canonicalize_github_repos(cap=5) and
ranked by relevance. The downstream block then re-canonicalized with
cap=None, which can re-sort by topic-slug match and clobber the
auto_resolve relevance order. Guard the second canonicalization with
a repos_from_auto_resolve flag so it only fires for user-supplied
--github-repo input.
Without this entry, the planner's _default_sources_for_intent() drops
xquik from the candidate pool for how_to / comparison / news intents
because SOURCE_CAPABILITIES.get("xquik", set()) returns the empty set.
Users with XQUIK_API_KEY set get zero Xquik results even though the
engine recognizes the key.
Mirrors the capabilities for "x" since both are X/Twitter-shaped
discussion + social sources.
Fixes#319
When yt-dlp is installed but stale (or otherwise unable to fetch transcripts
for any returned videos), runs previously reported YouTube as fully
successful in two user-facing surfaces:
1. Footer (render.py): showed "N videos | M views" with no indication
that zero transcripts were captured. The "with transcripts" segment
was conditionally suppressed when the count was zero - converting
the canonical stale-binary failure mode into a silent absence at
the very surface users read for "did this work?".
2. Quality nudge (quality_nudge.py): classified YouTube as "active"
based purely on yt-dlp installation + absence of a top-level error.
Per-video transcript-fetch ratio was never inspected. A run that
returned N videos with 0 transcripts (canonical stale-binary
failure) was reported as fully active.
The engine itself logs the failure correctly at default stderr level
(`[YouTube] Got transcripts for 0/N videos (N failed)`), but that line
gets buried in 100+ lines of parallel-source progress output and is
contradicted by the success-shaped footer and nudge that follow.
This change makes both conclusion surfaces honest:
* render.py footer always renders "M/N with transcripts" so the ratio
is visible regardless of value. Zero is no longer hidden. Format is
M/N (not bare M) so the denominator is in the message and the user
does not have to cross-reference the "videos" count.
* quality_nudge.py adds a third tier between "active" and "missing":
"degraded". Triggered when yt-dlp is installed AND videos were
returned AND transcript-fetch ratio is below threshold (default 50%,
tunable via DEGRADED_TRANSCRIPT_THRESHOLD env var). Emits an
actionable nudge: "YouTube returned N videos but only M transcripts
captured. The most common cause is a stale yt-dlp binary - YouTube's
caption format changes frequently and old binaries silently fail
every transcript. Update via your package manager: scoop update
yt-dlp (Windows), brew upgrade yt-dlp (macOS), or pip install -U
yt-dlp."
* last30days.py populates youtube_videos_count and
youtube_transcripts_count in the research_results dict it passes to
compute_quality_score, enabling the new degraded check at the call
site.
Threshold rationale: 50% accommodates a few legitimate
caption-disabled videos in a multi-video result, but a stale-binary
run that fails every transcript trips the nudge cleanly.
Score impact: degradation is informational, not score-affecting.
YouTube still counts as "active" in score_pct so users do not see
their score drop for a fixable client-side issue. The nudge directs
them to their own package manager.
Tests:
* tests/test_quality_nudge.py: 6 new TestYouTubeDegraded cases cover
zero-transcripts-flags-degraded, partial-above-threshold-does-not-flag,
zero-videos-does-not-flag (no false positives on absence),
one-of-three-flags-degraded, threshold-tunable-via-config, and
degraded-does-not-affect-score.
* tests/test_render_v3.py: 4 new YoutubeFooterTranscriptRatioTests
cases cover zero-transcripts-with-videos-renders-zero-over-total
(the regression repro), partial-renders-ratio, full-renders-ratio,
and no-videos-suppresses-entire-segment.
All 29 new test cases verified GREEN with the fix and RED without it
(temp-reverted both files separately to confirm each test catches the
specific regression it asserts).
Integration validation: ran the engine against an intentionally stale
yt-dlp 2025.03.31 binary placed first on PATH. Pre-fix the footer
showed `YouTube: 3 videos | 386,815 views` (no transcript signal).
Post-fix the footer shows `YouTube: 3 videos | 386,815 views | 0/3
with transcripts` and stderr emits "Degraded: YouTube" plus the
actionable update-yt-dlp nudge.
Out of scope (deserves its own PR): exposing transcripts_captured in
the EVIDENCE FOR SYNTHESIS block so the synthesizing model can flag
degradation in prose. Larger schema-touching change.
If CLAUDE_PLUGIN_ROOT ever expands to a path containing whitespace
(e.g. ~/Library/Application Support/...), the unquoted ${CLAUDE_PLUGIN_ROOT}
in hooks/hooks.json word-splits and bash receives the path as multiple
arguments, failing with "No such file or directory" on the first split.
Quoting the expansion makes the invocation correct regardless of the
characters in the resolved path. Verified manually:
unquoted + space -> bash: /tmp/with: No such file or directory
quoted + space -> bash: /tmp/with spaces/.../check-config.sh: No such file
quoted + real -> /last30days: Ready - 7 sources active.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Asserts compute_save_path_display() never returns a backslash when the
save_dir is under the user's home directory, regardless of host OS.
Reproduces the original bug on Windows (failed message before the fix:
AssertionError: '\' unexpectedly found in
'~/l30d_save_path__luu2g76\Documents\Last30Days\british-airways-middle-east-raw-v3.md'
)
and locks in the contract on POSIX hosts (passes trivially today; would
fail if anyone removes .as_posix() in the future).
Verified by temporarily reverting the fix and confirming RED, then
re-applying the fix and confirming GREEN. All 13 CliV3Tests pass.
The footer line `📎 Raw results saved to ~/Documents\Last30Days\…`
mangled the home-relative path on Windows because `f"~/{relative}"`
stringifies a `pathlib.Path` with the OS-native separator. The result
mixes a Unix tilde with backslashes, which neither File Explorer,
PowerShell, nor a `file://` URI can resolve.
`Path.as_posix()` always returns forward slashes, which is the
convention `~/`-prefixed paths require on every platform. macOS and
Linux output is unchanged because their separator is already `/`.
Repro on Windows:
python3 last30days.py "anything" --emit=compact --save-dir="$HOME/Documents/Last30Days"
# before: 📎 Raw results saved to ~/Documents\Last30Days\anything-raw.md
# after: 📎 Raw results saved to ~/Documents/Last30Days/anything-raw.md
Both files lived at the repo root as pre-plugin-layout artifacts. On
current main neither is referenced from README, SKILL.md, AGENTS.md,
CHANGELOG, or docs/ — no inbound links to break by removing. Git history
preserves the content for anyone who needs to dig it up.
Closes#352, #353. The PRs by @dinakars777 correctly flagged the drift;
deletion is the cleaner resolution than annotating them as historical.
macOS ships /bin/bash 3.2 and the script uses #!/bin/bash with
set -euo pipefail, so declare -g would abort the SessionStart hook
with "invalid option" on every Mac. printf -v writes via assignment
semantics (global from inside a function on 3.2+) — same scope
outcome, broader compatibility.
Trevin Chow