The instruction tells Copilot to verify PyPI/npm package and symbol names
against the attester.dev existence oracle (free keyless tier, 25 calls/day
per client IP) before installing or importing, catching hallucinated
dependency names (5.2-21.7% of LLM suggestions per USENIX Security 2025).
The preToolUse hook enforces the same check on code the agent writes:
stdlib-only script, blocks only on a confident oracle negative, fails open
on quota or network trouble, caches answers on disk. README tables
regenerated with npm start.
Sanitize URL input before embedding it in Copilot handoff prompts in both bash and PowerShell hook scripts to prevent command/prompt injection from untrusted link text.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* new hook fix-broken-links
* codespell: add ans for variable short for answer
* update: scripts hand off alternative url to copilot cmd
* codespell: add ext. arcade-canvas/game/phaser.min.js
* Apply suggestions from code review
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
* update: rm em-dash from hook scripts
---------
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
* Tool Guardian Hook
Add Tool Guardian hook for blocking dangerous tool operations
Introduces a preToolUse hook that scans Copilot agent tool invocations
against ~20 threat patterns (destructive file ops, force pushes, DB drops,
permission abuse, network exfiltration) and blocks or warns before execution.
* Address review feedback: move hook to .github/, remove accidental log file
- Move hooks/tool-guardian/ to .github/hooks/tool-guardian/
- Remove accidentally committed guard.log
- Update all path references in README.md
* Move log directory to .github/, revert hook files back to
hooks/
- Revert hook files from .github/hooks/ back to hooks/tool-guardian/
- Update default log path to .github/logs/copilot/tool-guardian/
- Update all path references in README.md and hooks.json
* New hook: secrets-scanner
Add a secrets-scanner hook that scans files modified during a Copilot
coding agent session for leaked secrets, credentials, and sensitive data.
The hook runs on sessionEnd and inspects files in one of three scopes:
- diff: only files changed in the current session (default)
- staged: only files currently staged in the git index
- all: every tracked file in the repository
Detected pattern categories:
- AWS access keys and secret keys
- GCP service account credentials
- Azure client secrets and storage connection strings
- GitHub personal access tokens
- Slack tokens (bot, user, webhook)
- Private key headers (RSA, EC, DSA, OpenSSH, PEM)
- Generic high-entropy bearer tokens
- Internal IP:port strings
Configurable via environment variables (SCAN_MODE, SCAN_SCOPE,
SECRETS_ALLOWLIST) so teams can tune for their workflow without
editing the script. Patterns are POSIX ERE (grep -E) compatible,
with no PCRE metacharacters, for portability across macOS and Linux.
Files: hooks.json, scan-secrets.sh, README.md
* refactor: move PATTERNS array to top of scan-secrets.sh for discoverability
Move the PATTERNS declaration to the top of the file so it is clearly
visible and easy to customize, as suggested in code review. Added a
descriptive header comment. No functional changes.
---------
Co-authored-by: Shehab Sherif <shehabsherif0@users.noreply.github.com>
- Switch from colon to tab delimiter to handle colons in evidence text
- Base64-encode evidence to prevent parsing issues
- Use MAX_SEVERITY in log output and JSON events
- Narrow regex patterns to reduce false positives:
- third[_-]?party instead of third.?party
- Role reassignment scoped to AI terms
- System prompt injection requires 'you are' context
- Fix session-end stats to scope to current session only
- Update privacy statement to clarify evidence snippets are logged
- Rename credential description to 'Possible hardcoded credential'
- Fix database destruction regex to also match semicolons
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Fix getResourceType() to match relative paths like hooks/<name>/README.md
and skills/<name>/SKILL.md using regex instead of string includes
- Extract hook events from hooks.json via parseHookMetadata() instead of
non-existent frontmatter.event field in plugin README generation
- Update AGENTS.md to describe hooks as folder-based (README.md + hooks.json)
instead of .hook.md files
- Update session-logger README to accurately reflect what scripts log
(remove references to sessionId, duration, prompt content)
- Introduced hooks to enable automated workflows triggered by specific events during GitHub Copilot sessions.
- Added documentation for hooks in AGENTS.md and README.md.
- Created a new directory structure for hooks, including README.md and hooks.json files.
- Implemented two example hooks: Session Auto-Commit and Session Logger.
- Developed scripts for logging session events and auto-committing changes.
- Enhanced validation and parsing for hook metadata.
- Updated build and validation scripts to accommodate new hooks functionality.
Mostafa Amini
Aaron Powell
John Haugabook
Hannah Gould
Ajith Raghavan
Shehab Sherif
Imran Siddique