unified-context-stack
Skill6 starsChanged 51 days agoMITView on GitHub
Use when any SDLC phase needs external docs, dependency internals, cross-session memory, or feature specs — provides tiered context retrieval across External Truth (docs), Internal Truth (dependencies), Historical Truth (memory), and Intent Truth (feature specs) with graceful degradation based on installed tools.
In this repository27 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- agent-safety-reviewSkill
- agent-team-executionSkill
- agent-team-reviewSkill
- alert-hygieneSkill
- authorial-judgmentSkill
- batch-scriptingSkill
- capture-knowledgeSkill
- deploy-gateSkill
- design-debateSkill
- implementation-drift-checkSkill
- improvement-minerSkill
- incident-analysisSkill
- incident-trend-analyzerSkill
- openspec-shipSkill
- outcome-reviewSkill
- panelSkill
- product-discoverySkill
- project-verificationSkill
- prototype-labSkill
- runtime-validationSkill
- second-opinionSkill
- security-scannerSkill
- skill-scaffoldSkill
- supply-chain-investigationSkill
- synthesizeSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Check the Context Stack: line from your session start for capability flags
- Distinctive: line 13
Read the relevant phase document for your current SDLC phase
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Unified Context Stack
An infrastructure-level skill that provides tiered context retrieval for every SDLC phase. Reads your session's Context Stack: capabilities line to determine which tools are available, then follows strict fallback tiers.
How to Use
- Check the
Context Stack:line from your session start for capability flags - Read the relevant phase document for your current SDLC phase
- For each capability dimension needed, follow the tier document in strict order
Capability Flags
These are injected by the session-start hook as: Context Stack: context7=true, context_hub_cli=false, ...
| Flag | Tool | What it enables |
context7 | Context7 MCP | Broad library doc retrieval |
context_hub_available | Context Hub via Context7 | High-trust curated docs — flag means Hub is reachable, not that it has docs for your library (query /andrewyng/context-hub) |
context_hub_cli | chub CLI | Local curated doc retrieval and annotations |
serena | Serena MCP | LSP-powered dependency mapping and AST edits |
serena_connected | Serena MCP | MCP server is connected (not just registered); set only when SERENA_CONNECTION_CHECK=1. false means NOT-CONNECTED or NOT-CHECKED — the probe only ever sets true, so the two are indistinguishable; treat false as unknown unless you set the env gate yourself |
forgetful_memory | Forgetful Memory | Persistent cross-session architectural knowledge |
forgetful_connected | Forgetful Memory | MCP server is connected (not just registered); set only when FORGETFUL_CONNECTION_CHECK=1. false means NOT-CONNECTED or NOT-CHECKED — the probe only ever sets true, so the two are indistinguishable; treat false as unknown unless you set the env gate yourself |
openspec | OpenSpec CLI | Whether the openspec binary is available. See the separate OpenSpec: capability line for detailed surface/command info. Intent Truth retrieval does NOT require this flag — it checks artifact presence directly. |
Tier Documents
- External Truth — API documentation retrieval
- Internal Truth — Blast-radius mapping and safe code edits
- Historical Truth — Institutional memory retrieval and storage
- Intent Truth — Feature specification and design rationale retrieval
Note: Intent Truth checks for artifact presence in the workspace (openspec/specs/, openspec/changes/, docs/plans/, docs/plans/archive/, docs/superpowers/specs/ [legacy]). The OpenSpec: capability line from session-start indicates CLI availability for write operations (used by openspec-ship), but Intent Truth retrieval works regardless of CLI installation — it reads local files.
Phase Documents
- Design — Intent and historical context before proposing approaches
- Triage & Plan — Context gathering before writing plans
- Implementation — Mid-flight lookups during execution
- Testing & Debug — Error resolution and live issue discovery
- Code Review — Claim verification and dependency checks
- Ship & Learn — Memory consolidation before session close
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.