🎩 You're Invited:Meet the Socket team at Black Hat in Las Vegas, August 3-6.RSVP
Sign In

agents-opencode

Package Overview
Dependencies
Maintainers
1
Versions
19
Alerts
File Explorer

Advanced tools

Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

agents-opencode - npm Package Compare versions

Comparing version
1.4.2
to
1.5.0
+6
-2
.opencode/agents/blogger.md

@@ -77,6 +77,10 @@ ---

1. Read `AGENTS.md` for project context and recent activity
2. Review any prior content creation patterns and topics covered
2. Read `state/session-state.json` for active goals/risks (if present)
3. Read `handoff/latest.md` for continuation context (if present)
4. Review any prior content creation patterns and topics covered
**At task completion:**
Update `AGENTS.md` with timestamped entry (latest first):
1. Update `state/session-state.json` with content decisions, risks, and next actions (if state file is in use).
2. Generate or refresh handoff packet using project tooling when content workflow phase changed.
3. Then update `AGENTS.md` with timestamped entry (latest first):

@@ -83,0 +87,0 @@ ```markdown

@@ -136,6 +136,10 @@ ---

1. Read `AGENTS.md` for project context and recent activity
2. Review any established patterns and conventions from prior sessions
2. Read `state/session-state.json` for active goals/risks (if present)
3. Read `handoff/latest.md` for continuation context (if present)
4. Review any established patterns and conventions from prior sessions
**At task completion:**
Update `AGENTS.md` with timestamped entry (latest first):
1. Update `state/session-state.json` with implementation decisions, open risks, and next actions.
2. Generate or refresh handoff packet using project tooling when implementation phase meaningfully changed.
3. Then update `AGENTS.md` with timestamped entry (latest first):

@@ -142,0 +146,0 @@ ```markdown

---
description: Engineering Manager advisor for leadership decisions, team dynamics, and technical strategy
mode: primary
temperature: 0.3
temperature: 0.25
steps: 30

@@ -135,5 +135,11 @@ permission:

**At session start:** Read `AGENTS.md` for project context and review existing EM-related documents.
**At session start:**
1. Read `AGENTS.md` for project context and review existing EM-related documents.
2. Read `state/session-state.json` for active goals/risks (if present).
3. Read `handoff/latest.md` for continuation context (if present).
**At task completion:** Update `AGENTS.md` with timestamped entry (latest first, 3-5 bullets max):
**At task completion:**
1. Update `state/session-state.json` with key recommendations, risks, and next actions.
2. Generate or refresh handoff packet using project tooling when advisory state changed.
3. Then update `AGENTS.md` with timestamped entry (latest first, 3-5 bullets max):

@@ -140,0 +146,0 @@ ```markdown

@@ -221,2 +221,22 @@ ---

## Progress Tracking for Long-Running Work
For complex or multi-phase tasks, include and maintain a status table in updates.
Use this format:
```markdown
## Workstream Status
| ID | Initiative | Impact / Effort | Status | Notes |
|---|---|---|---|---|
| S1 | [Initiative] | [High/Medium/Low] / [High/Medium/Low] | [✅ Done / 🔄 In Progress / ⏳ Planned / ⛔ Blocked] | [Short note] |
```
Update cadence:
- Include the table at plan start for long-running/complex tasks.
- Update status after each completed phase or loop cycle.
- Keep exactly one active item as `🔄 In Progress` where possible.
- Reflect blockers immediately with `⛔ Blocked` and mitigation options.
## Safety & Validation

@@ -242,6 +262,10 @@ - Verify each phase completes successfully

1. Read `AGENTS.md` for project context and recent activity
2. Apply successful orchestration patterns from previous sessions
2. Read `state/session-state.json` for working memory (if present)
3. Read `handoff/latest.md` for continuation context (if present)
4. Apply successful orchestration patterns from previous sessions
**At task completion:**
Update `AGENTS.md` with timestamped entry (latest first):
1. Refresh `state/session-state.json` with current phase, risks, decisions, and next actions.
2. Generate or refresh handoff packet using project tooling when phase state changed.
3. Then update `AGENTS.md` with timestamped entry (latest first):

@@ -248,0 +272,0 @@ ```markdown

@@ -181,6 +181,10 @@ ---

1. Read `AGENTS.md` for project context and recent activity
2. Review prior planning decisions and patterns
2. Read `state/session-state.json` for active goals/risks (if present)
3. Read `handoff/latest.md` for continuation context (if present)
4. Review prior planning decisions and patterns
**At task completion:**
Update `AGENTS.md` with timestamped entry (latest first):
1. Refresh `state/session-state.json` with planning decisions, risks, and recommended next actions.
2. Generate or refresh handoff packet using project tooling when plan state changed.
3. Then update `AGENTS.md` with timestamped entry (latest first):

@@ -187,0 +191,0 @@ ```markdown

@@ -23,2 +23,6 @@ #!/usr/bin/env node

const AGENT_DIR_LEGACY = 'agent';
const PROJECT_TEMPLATE_FILES = [
path.join('state', 'session-state.json'),
path.join('handoff', '.gitkeep'),
];

@@ -426,2 +430,40 @@ function resolveAgentDirectory(opencodeDir) {

function installProjectTemplateFiles(sourceDir, scope, paths, backupSession) {
if (scope !== 'project') {
return { installedCount: 0, skippedCount: 0 };
}
let installedCount = 0;
let skippedCount = 0;
for (const relativePath of PROJECT_TEMPLATE_FILES) {
const src = path.join(sourceDir, relativePath);
if (!fs.existsSync(src)) {
warning(`Project template source is missing: ${relativePath}`);
continue;
}
const dest = path.join(paths.rootDir, relativePath);
ensureDir(path.dirname(dest));
if (fs.existsSync(dest)) {
skippedCount += 1;
continue;
}
try {
if (backupSession) {
backupSession.backupFile(dest, relativePath);
}
} catch {
// backup not required for first-write path
}
fs.copyFileSync(src, dest);
installedCount += 1;
}
return { installedCount, skippedCount };
}
function mergeInstallerConfig(targetConfigPath, sourceConfig, onBeforeWrite) {

@@ -755,2 +797,3 @@ const patch = {

const {
sourceDir,
sourceConfig,

@@ -783,2 +826,12 @@ sourceOpencodeDir,

const templateResult = installProjectTemplateFiles(sourceDir, scope, paths, backupSession);
if (scope === 'project') {
if (templateResult.installedCount > 0) {
success(`✓ Installed ${templateResult.installedCount} project template file(s)`);
}
if (templateResult.skippedCount > 0) {
info(`Skipped ${templateResult.skippedCount} existing project template file(s)`);
}
}
if (languages) {

@@ -815,2 +868,10 @@ filterLanguages(paths.opencodeDir, languages);

const managedFiles = buildManagedFilesFromSource(scope, sourceManagedFiles, paths);
if (scope === 'project') {
for (const templateRelativePath of PROJECT_TEMPLATE_FILES) {
const absoluteTemplatePath = path.join(paths.rootDir, templateRelativePath);
if (fs.existsSync(absoluteTemplatePath)) {
managedFiles.push(templateRelativePath);
}
}
}
const manifest = {

@@ -1328,2 +1389,3 @@ schemaVersion: 1,

const ok = installScope({
sourceDir,
sourceConfig,

@@ -1366,2 +1428,3 @@ sourceOpencodeDir,

const ok = installScope({
sourceDir,
sourceConfig,

@@ -1381,2 +1444,3 @@ sourceOpencodeDir,

const ok = installScope({
sourceDir,
sourceConfig,

@@ -1383,0 +1447,0 @@ sourceOpencodeDir,

{
"name": "agents-opencode",
"version": "1.4.2",
"version": "1.5.0",
"description": "OpenCode Agents: Intelligent AI assistants for software development. Features 8 specialized agents, 14 coding standards, automated code review, documentation generation, and cross-platform installation. Supports .NET, Python, TypeScript, Flutter, Go, Java, Node.js, React, Ruby, and Rust with plan-first execution and context-aware assistance.",

@@ -23,4 +23,13 @@ "files": [

"validate:agents": "node scripts/validate-agents.js",
"validate:commands": "node scripts/validate-command-matrices.js",
"eval:agents": "node evals/harness/run-evals.js",
"eval:agents:json": "node evals/harness/run-evals.js --json --out evals/reports/latest.json",
"eval:agents:trend": "node scripts/compare-eval-trend.js --current evals/reports/latest.json --baseline evals/fixtures/eval-trend-baseline.json --out evals/reports/trend-summary.md",
"validate:installer": "node scripts/test-installer.js",
"validate:docs": "node scripts/validate-docs.js",
"validate:docs:external": "node scripts/validate-docs.js --external",
"validate:docs:external:allowlist": "node scripts/validate-docs.js --external --allowlist",
"validate:changelog": "node scripts/validate-changelog-labels.js",
"validate:session": "node scripts/validate-session-state.js",
"handoff:generate": "node scripts/generate-handoff.js --state state/session-state.json --out handoff/latest.md",
"validate:context": "node scripts/check-context-size.js",

@@ -27,0 +36,0 @@ "validate:all": "npm run doctor"

@@ -170,5 +170,20 @@ # OpenCode Agents

## Agent Evals
- `npm run eval:agents` runs deterministic contract checks for agent and command metadata.
- `npm run eval:agents:json` writes machine-readable output to `evals/reports/latest.json`.
- These checks are integrated into doctor and CI validation summary gating.
## Validation
- Run `npm run doctor` for the complete local validation suite.
- For full check mapping (local commands ↔ CI gates), see **[Compatibility](./docs/compatibility.md)**.
## Docs
- **[Getting Started](./docs/getting-started.md)**
- **[Approval Gates](./docs/approval-gates.md)**
- **[Compatibility](./docs/compatibility.md)**
- **[Deprecation & Migration Policy](./docs/deprecation-migration.md)**
- **[State Management](./docs/state-management.md)**
- **[Full Documentation](https://shahboura.github.io/agents-opencode/)**