skillshield good first issues (Python)

SG-003 flags benign text such as "escalate the ticket" as CRITICAL 36 minutes ago

bug help wanted good first issue

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

SARIF: report files that could not be read 36 minutes ago

enhancement help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Test the 50-snippet cap on line-number lookup 36 minutes ago

enhancement good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Support NO_COLOR and a --no-color option 36 minutes ago

enhancement good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Add --format json to skillshield rules 36 minutes ago

enhancement good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Test on Python 3.13 and 3.14 36 minutes ago

enhancement good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Replace the hand-typed "Tests: 291 passing" badge 36 minutes ago

documentation good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: switching off TLS certificate checks for installs and downloads 36 minutes ago

good first issue new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: data hidden in emoji variation selectors 36 minutes ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: one-line code that runs decoded or decompressed data 36 minutes ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: preloading a library into other programs (LD_PRELOAD, DYLD_INSERT_LIBRARIES) 36 minutes ago

good first issue new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: changing git settings so other commands run attacker code 36 minutes ago

good first issue new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Accept a reviewed finding with a reason (.skillshield-ignore) about 1 hour ago

enhancement help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

One documentation page per rule, linked from every finding about 1 hour ago

documentation help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Scan a skill from a GitHub link before installing it about 1 hour ago

enhancement help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

SG-001: negated Chinese instructions are flagged as overrides about 2 hours ago

help wanted false-positive

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Scan the skills already installed on this computer (--installed) 4 days ago

enhancement help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: risky agent settings in a project (hooks, MCP servers, auto-approval) 4 days ago

help wanted new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Folder scans exit 0 when a file is HIGH; only CRITICAL fails 7 days ago

bug good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: decode base64 and scan what is inside 7 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Add a pre-commit hook and let skillguard scan take several files 7 days ago

enhancement help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Line numbers missing for multi-line matches and repeated matches 7 days ago

bug help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: instruction override in languages other than English 7 days ago

help wanted good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: persona that denies being an AI 9 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: right-to-left override characters (Trojan Source) 10 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Roadmap: runtime guard that blocks what a skill should not do 12 days ago

help wanted roadmap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Add a performance budget test for every rule 12 days ago

enhancement good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Lethal trifecta (SG-011) switches off whenever the injection rule is evaded 12 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Roadmap: lock a skill after review and detect changes 12 days ago

help wanted roadmap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Roadmap: judge what a skill does, not how it is worded 12 days ago

help wanted roadmap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Roadmap: optional semantic check that runs locally 12 days ago

help wanted roadmap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Measure false positives on real, safe skills 12 days ago

enhancement help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Warn when a skill's name imitates a well-known skill or brand 13 days ago

enhancement help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: instruction to replace the skill with remote content 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Read SKILL.md frontmatter: rate injection in name and description higher, and check allowed-tools 13 days ago

enhancement help wanted

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: asking for a crypto wallet recovery phrase or private key 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: instructions hidden in Unicode tag characters (ASCII smuggling) 13 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: instructions fetched from a URL at run time 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: instructions that change how other tools run (line jumping) 13 days ago

help wanted detection-gap new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: telling the agent to hide or downgrade security findings 13 days ago

help wanted detection-gap new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: fake <SYSTEM> or <IMPORTANT> tags and developer-mode framing 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: writing instructions into agent memory or config files 13 days ago

help wanted detection-gap new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: install script from a paste site, pasted into a terminal 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: pip or npm pointed at an untrusted package registry 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: fake prerequisite that downloads a password-protected archive and runs it 13 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: base64 decoded and piped into a shell 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Flag compiled Python (.pyc) shipped in a skill 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Look inside archives and flag password-protected ones 13 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Scan scripts without an extension (shebang) and add PowerShell and batch files 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Files SkillShield cannot read are skipped silently and the skill is reported SAFE 13 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: installing a system service, cron job or login item 13 days ago

help wanted detection-gap new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

False positive: documentation that quotes an attack phrase is rated CRITICAL 13 days ago

help wanted false-positive

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: dormant instruction that fires on a trigger word 13 days ago

help wanted detection-gap new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

New rule: hidden side actions the user is not told about 13 days ago

help wanted detection-gap new-rule

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: code that posts environment variables to a URL 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: plain-English instruction to send secret files to a URL 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: data leaked through a markdown image URL 13 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: piping a downloaded script into a shell 13 days ago

good first issue detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Detection gap: reworded instruction override 13 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Normalise text before matching: zero-width characters and lookalike letters 13 days ago

help wanted detection-gap

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

One HIGH finding is reported as MEDIUM, so skillshield scan exits 0 13 days ago

bug good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

Bug report: SARIF from a directory scan uses file names, not paths 13 days ago

bug good first issue

obielin/skillshield

20

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

Python

#agent-skills#ai-safety#ai-security#anthropic#claude#clawhavoc#llm-security#mcp#prompt-injection#supply-chain

obielin/skillshield

Open Issues Need Help