issues Search Results · language:Edge language:TypeScript language:HTML language:Python language:JavaScript language:CSS
Filter by
20.2M results
Epic: E4 · Depends on: T0
Why: Enterprise adopters map tools to recognized frameworks. Mapping categories/patterns to OWASP LLM Top 10 +
agentic-threat taxonomies makes SkillSpector legible to security ...
docs
P1
barak-venos
例：以下は12行・判定回数14回と正しく判定されていない
ダイスログ
[メイン] sCCB =28 【回避】 (1D100 =28) ＞ 5 ＞ 決定的成功/スペシャル
[メイン] sCCB =14 【回避】 (1D100 =14) ＞ 24 ＞ 失敗
[メイン] sCCB =28 【回避】 (1D100 =28) ＞ 45 ＞ 失敗
[メイン] sCCB =28 / 2【回避】 (1D100 =28) ...
Taku-Taku-Taku
Epic: E4 · Depends on: T2
Why: Becoming the standard requires outside contributors adding patterns. A documented template + fixture convention
lowers the barrier.
Scope
- [ ] docs/adding-a-pattern.md: ...
docs
good first issue
P1
Epic: E2 · Depends on: T3
Why: Teams vetting many skills need aggregate views: scores over time, top risk categories, which skills to block.
Scope
- [ ] Static dashboard ingesting a directory of JSON ...
enhancement
integration
P2
Epic: E3 · Depends on: T12
Why: General scanners miss agent-specific risks (prompt injection, tool poisoning, excessive agency). A reproducible
head-to-head proves differentiation honestly, including ...
benchmark
P1
Context
Epic #1450. Phase 1 MVP. With the orientation contract defined (#1451), expose the note-independent re-entry surface GET
/api/companion/orientation. The two derivers it needs are already note-independent ...
agent:blocked
area:runtime
companion-ui
prio:high
type:task
RasmusTho
Epic: E3 · Depends on: T2
Why: Credibility requires reproducible numbers. A public labeled dataset lets the community measure precision/recall.
Scope
- [ ] Publish a labeled dataset (synthetic + sanitized ...
benchmark
P0
I want to fix this under gssoc 2026. Thank You
img width= 476 height= 594 alt= Image src=
https://github.com/user-attachments/assets/8a5f77d7-d599-4549-aad8-7e149345f72d /
type:bug
anubhab1601
Epic: E2 · Depends on: T3
Why: A clean skill may later pull a newly-vulnerable dependency or a maintainer may turn malicious. Re-scanning on
schedule + on advisories keeps verdicts current.
Scope
...
enhancement
integration
P2
Context
Once the classifier engine moves to @aiclarity/shared (DR-014), both MinSpec and ScroogeLLM read
.minspec/calibration.json. If both also write it (e.g. ScroogeLLM persisting its own tier corrections), ...
idea
inbox
P2
harvest316
Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip! Restrict your search to the title by using the in:title qualifier.
Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip! Restrict your search to the title by using the in:title qualifier.