Is sentry safe to install?

SaferSkills independently audited sentry (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.

CategoryWeightCategory scoreContribution

Securityprompt, exec, net, exfil, eval

35%

100

35.0 pts

Supply chainhash, typosquat, maintainer, lockfile

20%

100

20.0 pts

Maintenancestaleness, pinning, CI

15%

100

15.0 pts

TransparencySKILL.md, perms, README

15%

100

15.0 pts

Communityinstalls, verify, response

15%

100

15.0 pts

Findings & checks · 0 flagged

Securityscore 100 · 0 findings

✓—All security checks passedNo findings in this category for the latest scan.pass

Supply chainscore 100 · 0 findings

✓—All supply chain checks passedNo findings in this category for the latest scan.pass

Maintenancescore 100 · 0 findings

✓—All maintenance checks passedNo findings in this category for the latest scan.pass

Transparencyscore 100 · 0 findings

✓—All transparency checks passedNo findings in this category for the latest scan.pass

Communityscore 100 · 0 findings

✓—All community checks passedNo findings in this category for the latest scan.pass

The primary manifest — the file an agent reads to learn what this artifact does.

skills/sentry/SKILL.md · 1 file

skills/sentry/SKILL.md4.5 KB · Markdown

name: sentry

description: Query Sentry errors, triage issues, and check release health via REST API. Use when debugging exceptions, investigating crash reports, triaging error issues, checking release stability, or any error monitoring and issue tracking task. Also triggers on mentions of Sentry, errors, exceptions, crashes, stack traces, or error rates.

sentry

Query errors, investigate exceptions, track issues, and monitor release health via the Sentry REST API. API usage is referenced through Context7 MCP; only company-specific rules are documented here.

Description

Applicable scenarios: production error investigation, issue triage, release health checks, crash analysis.

Internal URL: https://sentry-us.example.com (self-hosted Sentry, API base path /api/0/)
Organization slug: sentry

Variable	Description	Required
`SENTRY_AUTH_TOKEN`	Organization Token (Settings -> Organization Tokens -> Create)	Yes

Authentication: -H "Authorization: Bearer $SENTRY_AUTH_TOKEN"

The current Organization Token only has Release-related permissions by default. To query Issues/Events, create an Internal Integration in Developer Settings and enable Issue & Event: Read, Project: Read, Organization: Read.

Rules

Project Structure (7 projects)

Project slug	Product	Platform	Description
`app-android`	Main App	Android	Primary consumer app (most frequent releases)
`app-ios`	Main App	iOS	Primary consumer app
`backend-service`	Backend	Java	Backend service (shares releases with consumer-service)
`consumer-service`	Backend	Java	Backend consumer service
`app2-android`	Sub-brand App	Android	Sub-brand app
`app2-ios`	Sub-brand App	iOS	Sub-brand app
`app3-ios`	Security App	iOS	Security app

Release Version Naming Convention

Three formats, distinguished by platform:

Platform	Format	Example
Android	`{major}.{minor}.{patch}({commit_hash})`	`2.85.0(b4f16c)`
iOS	`{bundle_id}@{version}+{build_number}`	`[email protected]+30713`
Backend	`{YYYYMMDD}@{COMMIT_HASH}`	`20260212@3A5AE26`

iOS Bundle ID Reference:

Bundle ID	Product
`com.example.app`	Main App
`com.example.guard`	Guard App
`com.example.app2`	Sub-brand App
`com.example.app3`	Security App

Release Frequency

app-android: Multiple builds per day (CI auto-reports)
app-ios: Several times per week
backend-service / consumer-service: Approximately daily deployments, version numbers include date (YYYYMMDD@HASH), both projects share the same release
app2-*: Lower frequency

Hard Rules

Look before you act: Write operations (resolve / ignore / assign) must be preceded by reviewing the stack trace and impact scope
PII protection: Error context may contain user data; never share raw stack traces externally or expose user information

Common Workflows

Investigate production errors: Query unresolved issues for a project -> sort by time -> view latest event stack trace -> locate code line -> provide fix suggestions
Post-deployment health check: Query latest release -> check for new issues -> confirm Crash Free Rate -> continue monitoring for 30min
Batch triage: Filter by project + environment -> review stack traces one by one -> resolve / ignore / assign
Release tracking: /api/0/organizations/sentry/releases/?per_page=20&sort=date -> check latest releases

Examples

Bad

# Resolving without investigation
User: Help me handle the errors on Sentry
AI: I've resolved all unresolved issues for you.  <- Did not review stack traces and impact scope

# PII leakage
AI: The stack trace shows user email [email protected] triggered the exception...  <- Should not expose user information

Good

# Post-deployment check
User: Backend just released a new version, check if there are new errors on Sentry
AI: Querying latest Release...
  1. Found 20260303@A1B2C3D, linked to backend-service and consumer-service
  2. Checking issues associated with this Release...
  3. No new issues found, recommend continuing to monitor for 30min

# Client crash investigation
User: app-android latest version 2.85.0 has user reports of crashes
AI: Querying app-android project Release 2.85.0 related issues...
  1. Found latest build 2.85.0(b4f16c)
  2. Viewing unresolved issues, sorted by affected user count
  3. Viewing top issue stack trace -> locating code

sentry— agent skill

Is sentry safe to install?

sentry

Description

Rules

Project Structure (7 projects)

Release Version Naming Convention

Release Frequency

Hard Rules

Common Workflows

Examples

Bad

Good

Audit the pieces. Scan the whole. Decide.