Install
$ agentstack add skill-tomzx-agents-developer-trust-profile ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Developer Trust Profile
Manages per-developer trust profiles that inform PR review behavior. Each profile lives in its own file and accumulates observations across reviews. Call this skill to view a profile, create one from scratch, or update it after a PR review.
Prerequisites
- GitHub username (
$1) of the developer - Optional:
--after-reviewto update after a review
Profile Location
~/.developer-trust/{github_username}.md
Trust Levels
| Level | Meaning | Effect on quick-pr-review | |-------|---------|--------------------------| | trusted | Developer consistently produces high-quality PRs | Standard checks, lean toward approval on borderline cases | | neutral | Unknown or mixed track record | Standard checks, default behavior | | cautious | History of issues, missed edge cases, or unclear PRs | Apply stricter interpretation of checks, flag marginal cases | | always_reject | Persistent quality/policy issues warranting manual review | Never auto-approve; always require human review |
Workflow
Ensure ~/.developer-trust is a git repo
|
Read existing profile (or start fresh)
|
Mode: view or update?
/ \
View Update
| |
Print profile Fetch PR metadata
|
Synthesize observations
(strengths, weaknesses,
patterns seen in diff)
|
Update profile sections
+ append review history row
|
Write profile file
|
git commit
Steps
1. Initialize repository
Ensure ~/.developer-trust exists and is a git repository:
mkdir -p ~/.developer-trust
git -C ~/.developer-trust init
git init is idempotent: safe to run even if the directory is already a repository.
2. Resolve paths and arguments
Parse $@:
GITHUB_USERNAME:$1- If
--after-reviewflag present, extract: REPO: next argument (owner/repo)PR_NUMBER: next argumentOUTCOME: next argument (approvedornot_approved)
PROFILE_PATH: ~/.developer-trust/{GITHUB_USERNAME}.md
3. Load existing profile
If PROFILE_PATH exists, read it. Extract current trust level, strengths, weaknesses, patterns, and review history. If it does not exist, initialize with defaults (trust level: neutral, all sections empty).
4a. View mode (no --after-review)
Print the profile contents and summarize:
- Trust level and reason
- Strengths count
- Weaknesses count
- Number of reviews in history
Stop here.
4b. Update mode (--after-review present)
Fetch PR context
gh pr view {PR_NUMBER} --repo {REPO} \
--json number,title,body,author,additions,deletions,changedFiles,labels,headRefName
gh pr diff {PR_NUMBER} --repo {REPO}
Synthesize observations
Based on the PR diff, title, description, and outcome, identify:
- Strengths: positive patterns (e.g., "well-structured tests", "clear commit messages", "good error handling", "thorough documentation")
- Weaknesses: problem areas (e.g., "missing test coverage", "unclear PR description", "mixed concerns in single commit", "ignores edge cases")
- PR patterns: recurring behaviors (e.g., "tends to include unrelated refactors", "consistently links issues", "PRs are atomic and focused")
Merge new observations with existing ones: add new items, update or reinforce existing ones (do not duplicate), and remove items that are contradicted by recent evidence.
Update trust level
Recalculate trust level based on the full review history and current observations:
- If the developer has a long track record of clean, approved PRs: consider upgrading to
trusted - If recurring issues appear across multiple reviews: consider downgrading to
cautious - If issues are severe or repeated despite feedback: consider
always_reject - Do not change trust level based on a single review alone unless it is egregious
Update the trust level reason to reflect the current evidence.
Append review history row
Add a new row to the Review History table:
| {ISO_DATE} | {REPO} | #{PR_NUMBER} | {Approved/Not approved} | {one-line observation} |
5. Write the profile
Write or overwrite PROFILE_PATH with the updated content using the format below.
6. Commit the change
Stage and commit the updated profile:
git -C ~/.developer-trust add {GITHUB_USERNAME}.md
git -C ~/.developer-trust commit -m "Update {GITHUB_USERNAME} trust profile ({OUTCOME}: {REPO}#{PR_NUMBER})"
For view-mode runs (step 4a), skip this step.
Profile Format
# Developer Trust Profile: {github_username}
_Last updated: {ISO_DATE}_
## Trust Level
**Level**: trusted | neutral | cautious | always_reject
**Reason**:
## Overview
## Strengths
-
-
## Weaknesses
-
-
## PR Patterns
-
-
## Review History
| Date | Repository | PR | Outcome | Notes |
|------|------------|----|---------|-------|
| {ISO_DATE} | {owner/repo} | #{number} | Approved | |
Example Usage
Scenario 1: View an existing profile
/developer-trust-profile alice
Reads ~/.developer-trust/alice.md and prints a summary. If no profile exists, reports that alice has no profile yet.
Scenario 2: Create a profile from scratch via first review
/developer-trust-profile bob --after-review acme/api 42 approved
No existing profile for bob. Fetches PR #42, synthesizes initial observations, creates a new profile with trust level neutral and the first review history entry.
Scenario 3: Update an existing profile
/developer-trust-profile alice --after-review acme/api 55 not_approved
Loads alice's profile, fetches PR #55 diff, merges new observations, appends review history row. Reconsiders trust level if patterns suggest a change is warranted.
Scenario 4: Developer with persistent issues
/developer-trust-profile carol --after-review acme/api 88 not_approved
Carol's profile shows 4 consecutive non-approvals, all due to missing tests. Updates trust level to cautious with reason "Consistently submits PRs without adequate test coverage despite prior feedback."
Useful Commands Reference
| Command | Description | |---|---| | mkdir -p ~/.developer-trust && git -C ~/.developer-trust init | Initialize the trust store as a git repo | | gh pr view --repo --json ... | Fetch PR metadata | | gh pr diff --repo | Fetch the PR diff | | git -C ~/.developer-trust add && git -C ~/.developer-trust commit -m "..." | Stage and commit a profile update |
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: tomzx
- Source: tomzx/agents
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.